P. Kalogeropoulos

Cardiovascular risk prediction based on home blood pressure measurement : The didima study

Objective Although home blood pressure (HBP) is being used increasingly in clinical practice, the evidence on its prognostic value is still limited. This study in the general population investigated the value of HBP compared to office measurements (OBP) in predicting cardiovascular risk. Subjects and methods In 1997 all adults of the Didima area in Greece were invited to participate in a cross-sectional study involving OBP (two visits) and HBP measurements (3 days). Incident cardiovascular morbidity and cause-specific mortality were assessed after 8.2 ± 0.2 years (mean ± SD). Average OBP and HBP were used in Cox regression analysis of fatal and non-fatal cardiovascular events with age, gender, history of cardiovascular disease, use of antihypertensive medication, smoking and diabetes as covariates. Results A total of 662 subjects were analysed (mean age at baseline 54.1 ± 17.6 years). During follow-up 78 deaths (42 cardiovascular) and 67 cardiovascular events (fatal and non-fatal) were documented. Unadjusted hazard ratios for cardiovascular events per 1 mmHg blood pressure increase were for HBP systolic 1.034 (P < 0.001) and diastolic 1.037 (P < 0.01) and for OBP systolic 1.035 (P < 0.001) and diastolic 1.021 (P = 0.07). After adjustment for all available cardiovascular risk predictors, only diastolic OBP remained significant. The addition of HBP in the models already including OBP did not significantly improve the predictive ability. White coat but not masked hypertensives were at high risk. Conclusions This study showed that both HBP and OBP are significant predictors of cardiovascular risk in the general population. However, no prognostic superiority of HBP compared to OBP has been demonstrated.

Prognostic value of home blood pressure measurement.

Although self-monitoring of blood pressure by patients at home is being widely used in clinical practice, the evidence on its prognostic value is still limited. Five long-term studies with nearly 60 000 patients/year have provided prognostic information for home blood pressure measurements. Differences exist among these studies regarding the population characteristics, the sample size and follow-up, the methodology and protocol for office and home blood pressure measurement and the adjustment procedure for other risk factors. All these studies, nevertheless, showed systolic home blood pressure to be a significant predictor of cardiovascular risk, and three of them also showed prognostic value of diastolic home blood pressure. Moreover, the prognostic value of home blood pressure appeared to be consistently superior to that of conventional office measurements. The prognostic significance of the white coat and the masked hypertension phenomena detected by home measurements were investigated in two studies, one in treated hypertensive patients and another in a general population sample. These studies showed that patients with white-coat phenomenon have similar cardiovascular risk as those with low office and home blood pressure, whereas the masked hypertension phenomenon is associated with high risk as in patients with uncontrolled hypertension. In conclusion, the available evidence suggests that home blood pressure has strong prognostic value, which appears to be superior to that of the conventional office measurements. More outcome studies on the prognostic value of home blood pressure, however, are needed.

Can validated wrist devices with position sensors replace arm devices for self-home blood pressure monitoring? A randomized crossover trial using ambulatory monitoring as reference.

BACKGROUND Electronic devices that measure blood pressure (BP) at the arm level are regarded as more accurate than wrist devices and are preferred for home BP (HBP) monitoring. Recently, wrist devices with position sensors have been successfully validated using established protocols. This study assessed whether HBP values measured with validated wrist devices are sufficiently reliable to be used for making patient-related decisions in clinical practice. METHODS This randomized crossover study compared HBP measurements taken using validated wrist devices (wrist-HBP, Omron R7 with position sensor) with those taken using arm devices (arm-HBP, Omron 705IT), and also with measurements of awake ambulatory BP (ABP, SpaceLabs), in 79 subjects (36 men and 43 women) with hypertension. The mean age of the study population was 56.7 +/- 11.8 years, and 33 of the subjects were not under treatment for hypertension. RESULTS The average arm-HBP was higher than the average wrist-HBP (mean difference, systolic 5.2 +/- 9.1 mm Hg, P < 0.001, and diastolic 2.2 +/- 6.7, P < 0.01). Twenty-seven subjects (34%) had a > or =10 mm Hg difference between systolic wrist-HBP and arm-HBP and twelve subjects (15%) showed similar levels of disparity in diastolic HBP readings. Strong correlations were found between arm-HBP and wrist-HBP (r 0.74/0.74, systolic/diastolic, P < 0.0001). However, ABP was more strongly correlated with arm-HBP (r 0.73/0.76) than with wrist-HBP (0.55/0.69). The wrist-arm HBP difference was associated with systolic ABP (r 0.34) and pulse pressure (r 0.29), but not with diastolic ABP, sex, age, arm circumference, and wrist circumference. CONCLUSIONS There might be important differences in HBP measured using validated wrist devices with position sensor vs. arm devices, and these could impact decisions relating to the patient in clinical practice. Measurements taken using arm devices are more closely related to ABP values than those recorded by wrist devices. More research is needed before recommending the widespread use of wrist monitors in clinical practice. American Journal of Hypertension doi:10.1038/ajh.2008.176American Journal of Hypertension (2008); 21, 7, 753-758. doi:10.1038/ajh.2008.176.

Seasonal variation in meteorological parameters and office, ambulatory and home blood pressure: predicting factors and clinical implications.

This study investigated the relationship between seasonal variations in blood pressure (BP) and the corresponding changes in meteorological parameters and weather-induced patients’ discomfort. Hypertensives on stable treatment were assessed in winter-1, summer and winter-2 with clinic (CBP), home (HBP) and 24-hour ambulatory BP (ABP). Discomfort indices derived from temperature, humidity and atmospheric pressure that reflected subjects’ discomfort were evaluated. Symptomatic orthostatic hypotension was assessed with a questionnaire. Sixty subjects (mean age 65.1±8.8 [s.d.], 39 men) were analyzed. CBP, HBP and daytime ABP were lower in summer than in winter (P<0.01). Nighttime ABP was unchanged, which resulted in a 55% higher proportion of non-dippers (P<0.001). All the discomfort indices that reflected weather-induced subjects’ discomfort were higher in summer (P<0.05) and systolic daytime ABP was <110 mm Hg in 15 subjects (25%). Seasonal changes in temperature and the discomfort indices were correlated with BP changes (P<0.05). Multivariate analyses revealed that winter BP levels, seasonal differences in temperature, female gender and the use of diuretics predicted the summer BP decline. In conclusion, all aspects of the BP profile, except nighttime ABP, are reduced in summer, resulting in an increased prevalence of non-dippers in summer with unknown consequences. Seasonal BP changes are influenced by changes in meteorological parameters, anthropometric and treatment characteristics. Trials are urgently needed to evaluate the clinical relevance of excessive BP decline in summer and management guidelines for practicing physicians should be developed.

Atrial Fibrillation Detection During 24-Hour Ambulatory Blood Pressure MonitoringNovelty and Significance: Comparison With 24-Hour Electrocardiography

This study assessed the diagnostic accuracy of a novel 24-hour ambulatory blood pressure (ABP) monitor (Microlife WatchBP O3 Afib) with implemented algorithm for automated atrial fibrillation (AF) detection during each ABP measurement. One hundred subjects (mean age 70.6±8.2 [SD] years; men 53%; hypertensives 85%; 17 with permanent AF; 4 paroxysmal AF; and 79 non-AF) had simultaneous 24-hour ABP monitoring and 24-hour Holter monitoring. Among a total of 6410 valid ABP readings, 1091 (17%) were taken in ECG AF rhythm. In reading-to-reading ABP analysis, the sensitivity, specificity, and accuracy of ABP monitoring in detecting AF were 93%, 87%, and 88%, respectively. In non-AF subjects, 12.8% of the 24-hour ABP readings indicated false-positive AF, of whom 27% were taken during supraventricular premature beats. There was a strong association between the proportion of false-positive AF readings and that of supraventricular premature beats (r=0.67; P<0.001). Receiver operating characteristic curve revealed that in paroxysmal AF and non-AF subjects, AF-positive readings at 26% during 24-hour ABP monitoring had 100%/85% sensitivity/specificity (area under the curve 0.91; P<0.01) for detecting paroxysmal AF. These findings suggest that in elderly hypertensives, a novel 24-hour ABP monitor with AF detector has high sensitivity and moderate specificity for AF screening during routine ABP monitoring. Thus, in elderly hypertensives, a 24-hour ABP recording with at least 26% of the readings suggesting AF indicates a high probability for AF diagnosis and should be regarded as an indication for performing 24-hour Holter monitoring.

DETECTION OF ATRIAL FIBRILLATION DURING ROUTINE 24-HOUR AMBULATORY BLOOD PRESSURE MONITORING IN THE ELDERLY: COMPARISON WITH 24-HOUR ELECTROCARDIOGRAPHY

Objective: To assess the diagnostic accuracy of a novel oscillometric 24-hour ambulatory blood pressure (ABP) monitor (Microlife WatchBP O3 Afib) with implemented algorithm for automated atrial fibrillation (AF) detection during each ABP measurement in elderly hypertensives. Design and method: Elderly subjects (age > 64 years, or 50–64 years, untreated or treated for hypertension, with stroke history or suspected arrhythmias) were subjected to simultaneous 24-hour ABP monitoring and 24-hour electrocardiography (ECG). Results: One hundred subjects (mean age 70.6 ± 8.2 [SD] years, men 53%, hypertensives 85%, 17 with permanent AF, 4 paroxysmal AF, 79 non-AF) were analyzed. Among a total of 6,410 valid ABP readings, 1,091 (17%) were taken in ECG AF rhythm. In reading-to-reading ABP analysis, the sensitivity, specificity and accuracy of the ABP monitor in detecting AF was 93%, 87% and 88% respectively. In non-AF subjects, 12.8% of the 24-hour ABP readings indicated false positive AF, of whom 27% were taken during supraventricular premature beats. There was a strong association between the proportion of false positive AF readings and that of supraventricular premature beats (r = 0.67, p < 0.001). In paroxysmal AF and non-AF subjects, the receiver operating characteristics curve (area under the curve 0.83, p < 0.01) showed that positive AF readings at 15% during 24-hour ABP monitoring had 90% sensitivity and 77% specificity for detecting AF (episode of any duration in 24-hour ECG). Conclusions: A novel 24-hour ABP monitor with AF detecting algorithm has high sensitivity and moderate specificity for AF screening during routine ABP monitoring in elderly hypertensives. AF detected in 15% or more of the 24-hour ABP readings in elderly hypertensives should mandate 24-hour ECG monitoring for suspected AF.

[PP.06.15] DETERMINANTS OF WHITE-COAT AND MASKED HYPERTENSION IN THE GENERAL POPULATION: THE DIDIMA STUDY

Objective: White-coat (WCH) and masked hypertension (MH) are common phenotypes between normotension and hypertension. This cross-sectional study investigated the clinical profile associated with increased likelihood for these phenomena. Design and method: Office (1 visit, 3 measurements) and home (3 days, 12 readings) blood pressure (BP) measurements were obtained in a general population study at Didima, Argolida, Greece. Average office and home BP were used to define WCH and MH. Participants’ characteristics (age, gender, BMI, diabetes and cardiovascular disease history, smoking, antihypertensive treatment) were assessed as potential determinants of WCH and MH. Results: 665 adults (age 54.4 ± 17.7 years, 42% men, 14.1% treated, 28.1% with office hypertension) were analyzed. The overall prevalence of WCH and MH was 8.7% and 7.5%, respectively (8.8% and 6.3% in untreated subjects, respectively). In multivariate logistic regression analysis, younger age (p = 0.004), untreated status (p = 0.044) and lower office BP (p = 0.01 for stage I versus stage II hypertension) were independent predictors of WCH among participants with office hypertension. For MH independent predictors were male gender (p = 0.034), ageing (p < 0.001), obesity (p < 0.001), antihypertension treatment (p = 0.03) and high-normal office BP (p = 0.01) among participants with office normotension. In sensitivity analysis performed in untreated subjects with office hypertension, only female gender (p = 0.048) and younger age (p = 0.001) increased the odds for WCH. In untreated subjects with low office BP older age (p < 0.001), higher BMI (p = 0.001), diabetes (p = 0.04) and high-normal office BP (p = 0.005) independently predicted MH. Figure. No caption available. Conclusions: In this general population study, age, gender, antihypertensive treatment and office BP level consistently predicted WCH and MH.

[OP.1C.07] PROGNOSTIC VALUE OF HOME VERSUS OFFICE BLOOD PRESSURE VARIABILITY: DIDIMA STUDY 19-YEAR FOLLOW-UP

Objective: Accumulating evidence suggests that blood pressure variability (BPV) predicts cardiovascular risk independently of the average blood pressure (BP). This study explored the prognostic ability of home versus office BPV. Design and method: In 1997 all adult population of Didima, Argolida, Greece was invited to participate in a study involving office (2 visits, 6 readings) and home (3 days, 12 readings) BP measurements. Cardiovascular morbidity and mortality were assessed after 19.0 ± 1.4 years (2016). Standard deviation (SD) and coefficient of variation (CV) of home and office BP were used to quantify BPV. Results: 665 participants (age 54.4 ± 17.7 years, 42% men) were analyzed. During follow-up 216 deaths (124 cardiovascular) and 146 cardiovascular events (fatal and non-fatal) were documented. In Cox regression models, all indices of systolic home BPV were predictive of cardiovascular risk, even after adjustment for all available cardiovascular risk factors (adjusted HR 1.05 for a 1-SD increase in systolic home BPV and 1.06 for 1% increment in percentage CV; all p < 0.05). The HR for SD (but not CV) of diastolic home BP significantly predicted cardiovascular outcome, yet it lost statistical significance after adjustment for several cardiovascular risk factors (adjusted HR 1.03 for a 1-SD increase in BPV, p = NS). For office BP, only unadjusted indices of BPV were predictive of cardiovascular risk (adjusted HR 1.01/1.03 for a 1-SD increase in systolic/diastolic BPV and 1.02/1.02 for 1% increment in percentage CV; all p = NS). In Cox model including CVs of both home and office systolic BP as independent variables, CV of home BP remained a significant predictor of cardiovascular outcome independent of office BPV. Figure. No caption available. Conclusions: In this long-term general population outcome study, both home and office BPV independently predicted cardiovascular events, with indices of home systolic BPV exhibiting superior prognostic ability.