CONGESTIVE HEART FAILURE

ORIGINAL CONTRIBUTION Diurnal Blood Pressure Pattern and Risk of Congestive Heart Failure Erik Ingelsson, MD, PhD Kristina Björklund-Bodegård, MD, PhD Lars Lind, MD, PhD Johan Ärnlöv, MD, PhD Johan Sundström, MD, PhD Context High blood pressure is the most important risk factor for congestive heart failure (CHF) at a population level, but the relationship of an altered diurnal blood pressure pattern to risk of subsequent CHF is unknown. Objectives To explore 24-hour ambulatory blood pressure characteristics as predictors of CHF incidence and to investigate whether altered diurnal blood pressure patterns confer any additional risk information beyond that provided by conventional office blood pressure measurements. Design, Setting, and Participants Prospective, community-based, observational cohort in Uppsala, Sweden, including 951 elderly men free of CHF, valvular disease, and left ventricular hypertrophy at baseline between 1990 and 1995, followed up until the end of 2002. Twenty-four-hour ambulatory blood pressure monitoring was performed at baseline, and the blood pressure variables were analyzed as predictors of subsequent CHF. Main Outcome Measure First hospitalization for CHF. Results Seventy men developed heart failure during follow-up, with an incidence rate of 8.6 per 1000 person-years at risk. In multivariable Cox proportional hazards models adjusted for antihypertensive treatment and established risk factors for CHF (myocardial infarction, diabetes, smoking, body mass index, and serum cholesterol level), a 1-SD (9 mm Hg) increase in nighttime ambulatory diastolic blood pressure (hazard ratio [HR], 1.26; 95% confidence interval [CI], 1.02-1.55) and the presence of nondipping blood pressure (night-day ambulatory blood pressure ratio 1; HR, 2.29; 95% CI, 1.16-4.52) were associated with an increased risk of CHF. After adjusting for office-measured systolic and diastolic blood pressures, nondipping blood pressure remained a significant predictor of CHF (HR, 2.21; 95% CI, 1.12-4.36 vs normal nightday pattern). Nighttime ambulatory diastolic blood pressure and nondipping blood pressure were also significant predictors of CHF after exclusion of all participants who had an acute myocardial infarction before baseline or during follow-up. Conclusions Nighttime blood pressure appears to convey additional risk information about CHF beyond office-measured blood pressure and other established risk factors for CHF. The clinical value of this association remains to be established in future studies. JAMA. 2006;295:2859-2866 www.jama.com CONGESTIVE HEART FAILURE (CHF) is one of the most common, costly, disabling, and deadly diseases. 1 It constitutes a huge burden on health services and accounts for 1% to 2% of the total health care costs in industrialized countries. 2 Once diagnosed as having CHF, patients have a1in3chance of dying within 1 year and a 2 in 3 chance of dying within 5 years. 1 The mortality associated with CHF exceeds that of most cancers, although recent reports suggest an improving prognosis. 3 The predominant causes of CHF are hypertension and coronary heart disease, and high blood pressure (BP) is suggested to be the most important risk factor for CHF at a population level. 4 Ambulatory BP monitoring provides information that is not obtained from conventional office-based BP measurement, such as mean BP over a 24-hour period and night-day patterns. Previous studies have established that 24-hour BP measurements are powerful predictors of cardiovascular morbidity and mortality independent of office-measured BP and other established cardiovascular risk factors. 5-8 However, no previous studies have examined 24-hour ambulatory BP as a predictor of incident CHF in persons free of CHF at baseline. Thus, the primary aim of the current study was to analyze 24-hour ambulatory BP characteristics as predictors of CHF incidence in a communitybased sample of elderly men, adjusting for antihypertensive treatment and traditional risk factors for CHF. The secondary aim was to investigate whether 24-hour ambulatory BP patterns confer any value regarding the risk of future CHF beyond that conveyed by conventional office-measured BP. METHODS Study Sample This study is based on the Uppsala Longitudinal Study of Adult Men cohort (http://www.pubcare.uu.se/ulsam/), a health investigation focused on identifying metabolic risk factors for cardiovascular disease. All 50-year-old men Author Affiliations: Departments of Public Health and Caring Sciences (Drs Ingelsson, Björklund-Bodegård, Ärnlöv, and Sundström) and Medical Sciences (Drs Lind and Sundström), Uppsala University, Uppsala, Sweden; and AstraZeneca Research and Development, Mölndal, Sweden (Dr Lind). Corresponding Author: Erik Ingelsson, MD, PhD, Department of Public Health and Caring Sciences, Geriatrics, Uppsala University, Uppsala Science Park, SE-751 85 Uppsala, Sweden (erik.ingelsson@pubcare.uu.se). 2006 American Medical Association. All rights reserved. (Reprinted) JAMA, June 28, 2006 Vol 295, No. 24 2859

living in Uppsala in 1970-1973 were invited to join the study, of which 82% (2322 men) participated in the investigation. 9 The cohort was reinvestigated 20 years later (baseline of the present study: 1991-1995). Of the 1681 available 70-year-old men invited to the follow-up investigation, 73% (1221 men) participated. For the present study, 1036 of these men had valid 24-hour ambulatory BP recordings and data on all covariates. Based on the hospital register, 12 further participants were excluded because of a previous diagnosis of CHF and 9 were excluded because of a previous diagnosis of valvular disease; 64 were excluded because of electrocardiographic left ventricular hypertrophy (ECG-LVH) at the baseline examination. Participants with ECG-LVH were excluded because there was evidence of a significant interaction between this variable and 24-hour ambulatory diastolic BP (DBP; P=.03) and nighttime ambulatory DBP (P=.03). Because the number of participants with ECG-LVH was low, the analyses were restricted to participants without it. Thus, 951 men were eligible for the present investigation. A subsample (n=819) excluding all participants with myocardial infarction at baseline or during follow-up was also examined. All participants provided written informed consent and the ethics committee of Uppsala University approved the study. Baseline Examinations Examinations performed when the participants were 70 years old included a medical examination, a questionnaire, blood sampling (after an overnight fast), supine office BP measurement, 24-hour ambulatory BP monitoring, anthropometric measurements, and lipid determinations as previously described. 8,10 Office-based BP was measured in the right arm with a sphygmomanometer using the appropriate cuff size. Two recordings were made to the nearest 2 mm Hg after a 10-minute supine rest, and the mean of the 2 measurements was used for the analyses. Systolic and diastolic BP were defined as Korotkoff phases I and V, respectively. Twenty-four-hour ambulatory systolic BP (SBP) and DBP were recorded using Accutracker II equipment (SunTech Medical Instruments Inc, Raleigh, NC). Blood pressure recordings were made every 20 or 30 minutes between 6 AM and 11 PM and every 20 or 60 minutes between 11 PM and 6 AM. Data were edited by omitting all readings presumed to be erroneous, including readings of 0, DBP readings of more than 170 mm Hg, SBP readings of more than 270 or less than 80 mm Hg, and all readings in which the difference between SBP and DBP was less than 10 mm Hg. The measurements were interpreted and the data edited by skilled laboratory technicians blinded to the study outcome. The Accutracker II device showed satisfactory accuracy and precision according to available documentation at the time of investigation in the early 1990s. 11 The coefficients of variation for 24-hour SBP and DBP were 6.8% and 5.5%, respectively, determined in a reproducibility study performed in 22 participants approximately 1 month after the baseline examination. Short, fixed clock time intervals were used, defining daytime as 10 AM to 8 PM and nighttime as midnight to 6 AM. This method is used to eliminate the retiring and rising periods, during which BP is subject to considerable variation; it also helps to diminish variations between different age groups and cultures. 12 Pulse pressure was estimated as the difference between SBP and DBP. Sustained hypertension was defined as office BP of 140/90 mm Hg or higher and daytime ambulatory BP of 135/85 mm Hg or higher; isolated ambulatory hypertension (or masked hypertension) was defined as office BP of less than 140/90 mm Hg and daytime ambulatory BP of 135/85 mm Hg or higher; and isolated office hypertension (or white-coat hypertension) was defined as office BP of 140/90 or higher and daytime ambulatory BP of less than 135/85 mm Hg. 13 Nocturnal BP decline was assessed by the ratio between mean nighttime and daytime SBP, and nondipping was defined as the ratio between nighttime and daytime SBP of 1 or higher; ie, an absence of nocturnal BP reduction, also called inverted dipping. There are several reasons for supporting this definition and the use of night-day BP ratios; first, the ratio depends less on the specific BP level than on the absolute nocturnal BP decrease; second, the ratios are normalized for daytime BP level; and third, a ratio of 1 approximates the 95th percentile of the distribution of nightday ratios in normotensive persons. 14 Furthermore, this most extreme form of circadian rhythm disturbance has been previously demonstrated to be associated with the highest risk of cardiovascular mortality. 15 The presence of diabetes at baseline was defined as fasting plasma glucose of 7.0 mmol/l (126 mg/dl) or higher or use of oral hypoglycemic agents or insulin. 16 The ECG-LVH was defined as high-amplitude R waves according to the revised Minnesota code, 17 together with a left ventricular strain pattern. 4 Coding of smoking was based on interview reports and medication data were based on questionnaire responses. At baseline, 310 participants were taking antihypertensive medications: 186 with monotherapy, 101 with 2 antihypertensive drugs, and 23 with 3 or more drugs. The presence of valvular disease (International Classification of Diseases, Ninth Revision [ICD-9] codes 394-397 and 424 or International Statistical Classification of Diseases, 10th Revision [ICD-10] codes I05-I08 and I34-I37) and prior myocardial infarction (ICD-9 code 410 or ICD-10 code I21) was assessed from the hospital discharge register. The precision of the myocardial infarction diagnosis in the Swedish hospital discharge register is high. 18 Follow-up and Outcomes Ninety-four men had a hospital discharge register diagnosis of CHF between the 70-year-old baseline and December 31, 2002. The ICD heart failure codes 428 (ICD-9) and I50 (ICD-10) 2860 JAMA, June 28, 2006 Vol 295, No. 24 (Reprinted) 2006 American Medical Association. All rights reserved.

and hypertensive heart disease with congestive heart failure code I11.0 (ICD-10) were considered to be possible CHF. The medical records from the relevant hospitalizations were reviewed by 2 physicians (E.I. and L.L.) who, blinded to the baseline data, classified the cases as definite, questionable, or miscoded. Cases included in the present study were considered definite. The classification relied on the definition proposed by the European Society of Cardiology 19 and the review process has been described in detail previously. 20 No participants were lost to follow-up. Statistical Analyses Analyses were defined a priori. Data are reported as mean (standard deviation) or number (percentage). Logarithmic transformation was performed to achieve normal distribution for the 24-hour pulse pressure, nighttime SBP, and nighttime pulse pressure variables. The prognostic values of a 1-SD increase for the continuous variables, or presence vs absence for the dichotomous variables, for CHF incidence were investigated with Cox proportional hazards analyses. The different hypertension definitions (sustained hypertension, isolated ambulatory hypertension, and isolated office hypertension) were compared with the reference level with lowest CHF risk (ie, no hypertension, office BP 140/90 mm Hg and daytime ambulatory BP 135/85 mm Hg) using Cox proportional hazards analyses. Nonlinear relationships were assessed by examining incidence rates in quintiles of the independent variables, and night-day ratio of SBP was nonlinearly related to CHF incidence, supporting the use of a dichotomous nondipping variable. Proportional hazards assumptions were confirmed by Schoenfeld tests. We investigated the independent variables in 5 sets of models in a hierarchical fashion: 1. Unadjusted; 2. Models adjusted for antihypertensive medication (diuretics, -blockers, angiotensin-converting enzyme inhibitors, calcium antagonists, and -blockers as separate covariates); Table 1. Baseline Characteristics of the Cohort by Development of Heart Failure During Follow-up* Characteristics Participants Who Developed CHF (n = 70) Participants Who Did Not Develop CHF (n = 881) Prior myocardial infarction 10 (14) 53 (6) Diabetes prevalence 12 (17) 83 (9) Current cigarette smoking 25 (36) 169 (19) Body mass index, mean (SD) 27.0 (3.2) 26.1 (3.3) Serum cholesterol level, mean (SD), mmol/l 5.8 (0.8) 5.8 (1.0) Medications Diuretics 18 (26) 85 (10) -Blockers 24 (34) 159 (18) ACE inhibitors 8 (11) 38 (4) Calcium antagonists 19 (27) 95 (11) -Blockers 1 (1) 12 (1) Sustained 39 (56) 429 (49) Isolated ambulatory 9 (13) 92 (10) Isolated office 14 (20) 151 (17) Nondipping 11 (16) 45 (5) Abbreviations: ACE, angiotensin-converting enzyme; CHF, congestive heart failure. SI conversion: To convert cholesterol to mg/dl, divide by 0.0259. *Data are expressed as No. (%) of participants unless otherwise noted. Defined as weight in kilograms divided by height in meters squared. Defined as night-day ambulatory blood pressure ratio of at least 1. 3. Models adjusted for antihypertensive medication and established risk factors for CHF (prior acute myocardial infarction, diabetes, smoking, body mass index, and serum cholesterol) determined at baseline; 4. Models adjusted for antihypertensive medication, established risk factors for CHF, and office-measured SBP and DBP; and 5. Models adjusted for antihypertensive medication, established risk factors for CHF, and 24-hour ambulatory SBP and DBP. Models 2, 3, and 4 were considered to be primary, whereas models 1 and 5 were considered secondary. All models were repeated in a subsample (n=819) without myocardial infarction before baseline or during followup. To rule out an effect modification by established risk factors on the relation of ambulatory BP variables to CHF, we investigated interaction terms between each of the established risk factors (those listed for model 3, ECG-LVH, and interim myocardial infarction) and the different BP variables. The statistical power was 91% to detect a hazard ratio (HR) of 1.5 for a 1-SD increase in BP level. Two-tailed 95% confidence intervals (CIs) and P values are reported, with P.05 regarded as statistically significant. PASS 2002 statistical software (NCSS, Kaysville, Utah) was used for power calculations and STATA version 8.2 (Stata Corp, College Station, Tex) for the other analyses. RESULTS Participants had a median follow-up time of 9.1 years (range, 0.1-11.4 years), contributing to 8129 person-years at risk. Seventy participants developed CHF during follow-up, with an incidence rate of 8.6 per 1000 personyears at risk. TABLE 1 shows the clinical characteristics at baseline and TABLE 2 shows the baseline BP characteristics. The frequency distribution of the ratio between mean nighttime and daytime SBP in the total cohort is shown in FIGURE 1. The mean (SD) nighttimedaytime SBP ratio was 0.85 (0.10). In unadjusted Cox proportional hazards analyses (model 1), all office BP measurements, 24-hour ambulatory BP measurements, nighttime ambulatory BP measurements, sustained hyperten- 2006 American Medical Association. All rights reserved. (Reprinted) JAMA, June 28, 2006 Vol 295, No. 24 2861

Table 2. Baseline Blood Pressure Measurements of the Total Cohort and by Development of CHF During Follow-Up Mean (SD), mm Hg Participants Who Developed CHF (n = 70) Participants Who Did Not Develop CHF (n = 881) Total Cohort (N = 951) Measurement SBP DBP PP SBP DBP PP SBP DBP PP Office 152 (19) 86 (10) 66 (15) 146 (18) 83 (9) 62 (14) 146 (18) 84 (9) 63 (14) 24-hour ambulatory 136 (12) 76 (7) 60 (10) 132 (16) 75 (8) 58 (12) 133 (16) 75 (8) 58 (12) Daytime ambulatory 143 (14) 80 (9) 63 (11) 140 (16) 79 (8) 60 (12) 140 (16) 79 (8) 61 (12) Nighttime ambulatory 124 (19) 69 (10) 55 (12) 118 (18) 66 (8) 52 (14) 119 (18) 66 (9) 52 (14) Abbreviations: CHF, congestive heart failure; DBP, diastolic blood pressure; PP, pulse pressure; SBP, systolic blood pressure. Figure 1. Frequency Distribution of the Ratio Between Mean Nighttime and Daytime Systolic Blood Pressure in the Overall Cohort Frequency, No. 150 125 100 75 50 25 0 0.6 0.8 1.0 1.2 1.4 Ratio Between Nighttime and Daytime Systolic Blood Pressure Fifty-six participants exhibited nondipping (a nighttime-daytime systolic blood pressure ratio 1). The blue curve is the kernel-density estimate using a bandwidth of 0.03. sion, and nondipping were significant predictors of CHF (TABLE 3). In analyses adjusted for antihypertensive treatment (model 2), nighttime ambulatory SBP and DBP and nondipping were significant predictors of CHF incidence (Table 3). When we also adjusted for established baseline risk factors for CHF (prior acute myocardial infarction, diabetes, smoking, body mass index, and serum cholesterol; model 3), nighttime ambulatory DBP and nondipping remained independent predictors of CHF in separate models (Table 3). A Kaplan-Meier plot for probability of survival free of heart failure for nondipping vs normal night-day BP patterns is presented in FIGURE 2. When office SBP and DBP were added to the models (model 4), nondipping was a significant predictor of CHF, whereas ambulatory nighttime DBP became nonsignificant (TABLE 4). In analyses including 24-hour ambulatory SBP and DBP in addition to antihypertensive treatment and established baseline risk factors (model 5), nighttime ambulatory DBP and nondipping were significant predictors of CHF (Table 4). To facilitate interpretation of the results, we also calculated how much the risk for CHF increased for each 5 mm Hg increment in nighttime ambulatory DBP. In unadjusted models (model 1), each 5 mm Hg increment of nighttime ambulatory DBP was associated with a 21% increased risk of CHF (HR, 1.21; 95% CI, 1.07-1.36). Adjusting for potential confounders in 4 different models, a 5 mm Hg increase in nighttime ambulatory DBP was associated with a 13% to 25% increased risk of CHF (HR, 1.16; 95% CI, 1.02-1.31 in model 2; HR, 1.14; 95% CI, 1.01-1.29 in model 3; HR, 1.13; 95% CI, 0.98-1.30 in model 4; and HR, 1.25; 95% CI, 1.01-1.54 in model 5). An increased nighttime ambulatory DBP, as well as a nondipping BP, were also associated with an increased absolute risk of CHF. The incidence of CHF was 2.7 cases higher per 1000 person-years at risk for each 5 mm Hg increment of nighttime ambulatory DBP (ranging from approximately 6 cases per 1000 person-years at risk in the lowest BP groups [50-65 mm Hg] to approximately 30 cases per 1000 person-years at risk in the highest groups [90-100 mm Hg]) and 15.1 cases higher per 1000 person-years at risk for those with nondipping vs normal nightday BP pattern (22.8 vs 7.7 cases). In the subsample without myocardial infarction before baseline or during follow-up, several BP variables were significant predictors of CHF in unadjusted analyses (model 1; Table 3). Adjusting for antihypertensive treatment (model 2) in this subsample, officemeasured SBP, 24-hour ambulatory DBP, nighttime ambulatory SBP and DBP, and nondipping were significant predictors of CHF incidence (Table 3). After further adjustment for the established risk factors for CHF (model 3), office SBP, nighttime ambulatory DBP, and nondipping remained significant predictors of CHF (Table 3). When adding office SBP and DBP to the models (model 4), nighttime ambulatory DBP and nondipping were significant predictors of CHF (Table 4). In analyses including 24-hour ambulatory SBP and 2862 JAMA, June 28, 2006 Vol 295, No. 24 (Reprinted) 2006 American Medical Association. All rights reserved.

Table 3. Risk of Congestive Heart Failure by Type of Blood Pressure Measurement in the Total Sample and in a Subsample Without Myocardial Infarction Hazard Ratio (95% Confidence Interval)* Unadjusted (Model 1) Adjusted for Antihypertensive Treatment (Model 2) Adjusted for Antihypertensive Treatment and Established Risk Factors (Model 3) Total sample (N = 951) Office measurements SBP 1.40 (1.11-1.75) 1.21 (0.95-1.53) 1.25 (0.98-1.59) DBP 1.31 (1.04-1.66) 1.13 (0.89-1.44) 1.16 (0.91-1.49) PP 1.30 (1.03-1.63) 1.17 (0.93-1.48) 1.20 (0.95-1.52) 24-hour ambulatory measurements SBP 1.26 (1.03-1.56) 1.17 (0.95-1.45) 1.13 (0.91-1.40) DBP 1.26 (1.01-1.58) 1.17 (0.93-1.47) 1.13 (0.90-1.42) PP 1.25 (1.00-1.55) 1.18 (0.95-1.47) 1.14 (0.91-1.42) Daytime ambulatory measurements SBP 1.20 (0.96-1.49) 1.13 (0.90-1.41) 1.08 (0.85-1.36) DBP 1.11 (0.88-1.40) 1.03 (0.81-1.31) 0.99 (0.78-1.26) PP 1.18 (0.95-1.46) 1.14 (0.92-1.41) 1.10 (0.88-1.37) Nighttime ambulatory measurements SBP 1.34 (1.09-1.64) 1.25 (1.01-1.54) 1.21 (0.98-1.49) DBP 1.38 (1.12-1.71) 1.29 (1.04-1.59) 1.26 (1.02-1.55) PP 1.25 (1.02-1.53) 1.17 (0.95-1.45) 1.14 (0.93-1.41) Sustained 2.27 (1.06-4.86) 1.74 (0.80-3.77) 1.75 (0.80-3.85) Isolated ambulatory 2.57 (0.99-6.67) 2.43 (0.94-6.31) 2.18 (0.83-5.76) Isolated office 2.25 (0.94-5.37) 1.81 (0.75-4.35) 2.01 (0.82-4.91) Nondipping 2.88 (1.51-5.48) 2.68 (1.38-5.21) 2.29 (1.16-4.52) Subsample without myocardial infarction (n = 819) Office measurements SBP 1.51 (1.13-2.03) 1.40 (1.03-1.91) 1.39 (1.02-1.89) DBP 1.36 (1.00-1.84) 1.21 (0.88-1.67) 1.21 (0.87-1.67) PP 1.42 (1.06-1.89) 1.35 (1.00-1.83) 1.35 (1.00-1.84) 24-ambulatory measurements SBP 1.37 (1.07-1.76) 1.27 (0.98-1.65) 1.19 (0.92-1.54) DBP 1.52 (1.16-2.00) 1.42 (1.07-1.88) 1.32 (0.99-1.75) PP 1.25 (0.95-1.66) 1.18 (0.89-1.56) 1.13 (0.86-1.49) Daytime ambulatory measurements SBP 1.28 (0.97-1.69) 1.20 (0.92-1.58) 1.13 (0.86-1.50) DBP 1.27 (0.96-1.69) 1.20 (0.90-1.61) 1.12 (0.83-1.50) PP 1.18 (0.90-1.56) 1.14 (0.87-1.49) 1.10 (0.83-1.44) Nighttime ambulatory measurements SBP 1.45 (1.13-1.86) 1.35 (1.04-1.75) 1.29 (0.99-1.66) DBP 1.66 (1.29-2.14) 1.53 (1.19-1.97) 1.48 (1.14-1.92) PP 1.24 (0.96-1.61) 1.16 (0.88-1.52) 1.13 (0.87-1.47) Sustained 3.27 (1.15-9.34) 2.69 (0.93-7.77) 2.52 (0.86-7.39) Isolated ambulatory 1.88 (0.42-8.40) 1.76 (0.39-7.88) 1.61 (0.35-7.31) Isolated office 2.28 (0.67-7.78) 1.93 (0.56-6.70) 2.15 (0.61-7.51) Nondipping 3.17 (1.41-7.15) 2.99 (1.28-7.00) 2.82 (1.20-6.62) Abbreviations: DBP, diastolic blood pressure; PP, pulse pressure; SBP, systolic blood pressure. *Cox proportional hazards ratios for a 1-SD increase in continuous variables or presence vs absence of dichotomous variables. The hazard ratios presented for sustained hypertension, isolated ambulatory hypertension, and isolated office hypertension were compared with a referent of no hypertension. Hazard ratios and 95% confidence intervals shown are unadjusted (model 1), adjusted for antihypertensive treatment (model 2), or adjusted for antihypertensive treatment and established risk factors for CHF (prior acute myocardial infarction, diabetes, smoking, body mass index, and serum cholesterol) determined at baseline (model 3). P.05 was considered statistically significant. P.01. P.05. P.001. Without myocardial infarction before baseline or during follow-up. Defined as night-day ambulatory blood pressure ratio of at least 1. 2006 American Medical Association. All rights reserved. (Reprinted) JAMA, June 28, 2006 Vol 295, No. 24 2863

Figure 2. Kaplan-Meier Plot for Probability of Heart Failure Free Survival in the Total Cohort for Nondipping vs Normal Night-Day Blood Pressure Patterns Probability of Heart Failure Free Survival No. at Risk Normal Night-Day Pattern Nondipping Pattern 1.0 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 0 DBP in addition to antihypertensive treatment and established baseline risk factors (model 5), nighttime ambulatory DBP and nondipping were significant predictors of CHF in this subsample (Table 4). None of the investigated interaction terms was significant in the main analysis sample (excluding men with prevalent CHF, valvular disease, or ECG-LVH) or in the sample without interim myocardial infarction. COMMENT In this community-based sample of elderly men free of CHF, valvular disease and ECG-LVH at baseline, a nondipping night-day BP pattern, and increased nighttime diastolic BP predicted CHF incidence independent of antihypertensive treatment and established risk factors for CHF, including myocardial infarction during the followup. Furthermore, a nondipping nightday BP pattern increased the risk of CHF even after adjusting for conventional office BP measurement. This indicates that nighttime BP patterns may be important in development of CHF and that a traditional office BP measurement does not capture all of the increased risk that an increased nighttime BP conveys. Twenty-four-hour ambulatory BP has repeatedly been demonstrated to be a Normal Night-Day Pattern Nondipping Pattern Log Rank P<.001 2 4 6 8 10 Follow-up, y 895 872 838 792 690 335 56 55 52 50 40 18 Nondipping pattern defined as night-day ambulatory blood pressure ratio of at least 1. powerful predictor of future cardiovascular morbidity and mortality, even after adjustment for conventional office BP. 5-8 Staessen et al 5 demonstrated that a nondipping BP pattern was associated with an increased cardiovascular risk and that nighttime BP more accurately predicted cardiovascular events than daytime BP. 5 Other studies also have identified a nondipping pattern as a risk factor for cardiovascular disease. 21,22 Previous studies have included CHF as a part of a combined end point but have not examined whether a 24-hour ambulatory BP measurement predicts CHF per se. The pathophysiology of atherosclerotic disease and CHF are not the same, and most new-onset CHF is not preceded by myocardial infarction. 4,23,24 Therefore, studies of predictors of CHF, accounting properly for myocardial infarction and risk factors for atherosclerotic disease, are warranted. A reduced circadian BP variation is a common finding in CHF patients, as reviewed by Goyal et al. 25 An increased nighttime ambulatory BP has been related to left ventricular filling impairment in cross-sectional studies with limited samples. 26,27 In a recent crosssectional study of patients with hypertension and type 2 diabetes mellitus, diastolic dysfunction was closely related to increased diastolic BP and nondipping. 28 However, to our knowledge, there are no previous studies of 24-hour ambulatory BP patterns as predictors of incident CHF. Several studies have established that nondipping is associated with endothelial dysfunction and hemostasis. 29-31 Recent studies indicate that endothelial dysfunction is an important component of the pathophysiological mechanisms of CHF, and it has been associated with the progression and prognosis of CHF. 32,33 Thus, it is possible that endothelial dysfunction could be a link between increased nocturnal BP and CHF. Another possible common pathophysiological mechanism is increased sympathetic activity, which is associated with a nondipping BP pattern 34,35 and also is a presumed causal factor in CHF. 36 One point of controversy is whether the absence of a nighttime BP decrease per se or an increased 24-hour BP load causes organ damage. To address this possibility, 24-hour ambulatory SBP and DBP were included as covariates in addition to antihypertensive treatment and established risk factors (model 5) in both samples. In these analyses, nondipping and nighttime DBP remained significant predictors of CHF, indicating that the nondipping BP pattern per se is important or is an indicator of an important trait. An example of such a trait could be sleep apnea, a condition that has been suggested to be associated both with CHF 37,38 and a nondipping BP pattern. 39 However, since this study was not designed to address the possible involvement of sleep apnea, this needs to be examined in future studies. A possible explanation for observed importance of nighttime ambulatory BP might be that the intraindividual BP variation is lower than that for daytime BP. This may be due to high BP consistency during sleep compared with daytime BP, which is influenced more by physical and psychological activity. However, it should be noted that this study did not primarily address the pathophysiological mechanisms behind the association between night- 2864 JAMA, June 28, 2006 Vol 295, No. 24 (Reprinted) 2006 American Medical Association. All rights reserved.

time BP pattern and CHF. That has to be examined in other settings. The strengths of this study include the large, community-based population and the long follow-up period. Furthermore, all CHF cases were validated, limiting the inclusion of false-positive cases. However, there are some limitations to this study. Because we only examined men of the same age with a similar ethnic background, this study has unknown generalizability to women or other age and ethnic groups. However, we did circumvent the powerful effects of age on CHF incidence. Moreover, due to limitations in sample size, men using antihypertensive medications were included in the study population, which may have affected the results through residual confounding, despite the adjustment for specific antihypertensive treatment. Another possible limitation is that multiple statistical testing is present to some degree. However, all analyses were specified a priori and the findings were consistent in all models and in the subsample. Since evidence of effect modification between some of the BP variables and ECG-LVH was found, the study sample had to be stratified into 2 parts: one stratum with participants with ECG-LVH at baseline and the other without these participants. The stratum consisting of participants with ECG-LVH was unfortunately too small (64 participants, of whom 11 developed CHF during the follow-up) to be analyzed. Thus, the analyses were restricted to the stratum with participants without ECG-LVH, which also might be considered to be a limitation of the study. Milder, nonhospitalized cases of CHF were not included in our end point, which may be considered a limitation but would tend to bias the results toward the null hypothesis. Since the CHF diagnosis was based on a review of medical records, it was not possible to differentiate between systolic and diastolic heart failure because echocardiography was not available at the time of diagnosis for many of the cases. Thus, we could not examine whether the impact of BP pattern is different for systolic vs diastolic heart failure. Finally, the diagnosis of Table 4. Risk of Congestive Heart Failure by Type of Blood Pressure Measurement in the Total Sample and in a Subsample Without Myocardial Infarction, Adjusted for Previous Factors Plus Office or Ambulatory Blood Pressure Hazard Ratio (95% Confidence Interval)* Adjusted for Adjusted for Antihypertensive Treatment, Established Risk Factors, and Office Blood Pressure (Model 4) Antihypertensive Treatment, Established Risk Factors, and 24-hour Ambulatory Blood Pressure (Model 5) Total sample (N = 951) Office measurements SBP NA 1.23 (0.92-1.65) DBP NA 1.11 (0.82-1.51) PP NA 1.19 (0.91-1.56) 24-hour ambulatory measurements SBP 1.01 (0.77-1.32) NA DBP 1.05 (0.79-1.39) NA PP 1.05 (0.80-1.37) NA Daytime ambulatory measurements SBP 0.94 (0.70-1.25) 0.78 (0.45-1.36) DBP 0.87 (0.66-1.16) 0.53 (0.30-0.92) PP 1.01 (0.78-1.30) 1.14 (0.65-2.02) Nighttime ambulatory measurements SBP 1.14 (0.89-1.44) 1.41 (0.92-2.14) DBP 1.23 (0.97-1.58) 1.47 (1.03-2.11) PP 1.07 (0.84-1.35) 1.24 (0.80-1.90) Sustained 0.98 (0.35-2.76) 1.41 (0.55-3.59) Isolated ambulatory 2.01 (0.76-5.33) 1.91 (0.69-5.29) Isolated office 1.32 (0.48-3.62) 1.95 (0.80-4.76) Nondipping 2.21 (1.12-4.36) 2.19 (1.06-4.53) Subsample without myocardial infarction (n = 819) Office measurements SBP NA 1.32 (0.91-1.92) DBP NA 1.01 (0.67-1.51) PP NA 1.37 (0.97-1.94) 24-hour ambulatory measurements SBP 1.04 (0.76-1.44) NA DBP 1.29 (0.90-1.84) NA PP 0.97 (0.70-1.36) NA Daytime measurements SBP 0.95 (0.67-1.35) 0.66 (0.31-1.38) DBP 1.00 (0.71-1.43) 0.45 (0.23-0.87) PP 0.94 (0.68-1.31) 1.12 (0.52-2.41) Nighttime measurements SBP 1.18 (0.88-1.58) 1.53 (0.88-2.65) DBP 1.47 (1.10-1.97) 1.62 (1.07-2.46) PP 1.00 (0.73-1.37) 1.21 (0.65-2.24) Sustained 1.65 (0.42-6.50) 1.97 (0.57-6.83) Isolated ambulatory 1.52 (0.33-6.97) 1.37 (0.29-6.55) Isolated office 1.55 (0.38-6.32) 2.11 (0.60-7.41) Nondipping 2.58 (1.09-6.12) 2.73 (1.10-6.76) Abbreviations: DBP, diastolic blood pressure; NA, not applicable; PP, pulse pressure; SBP, systolic blood pressure. *Cox proportional hazards ratios for a 1-SD increase in continuous variables or presence vs absence of dichotomous variables. The hazard ratios presented for sustained hypertension, isolated ambulatory hypertension, and isolated office hypertension were compared with a referent of no hypertension. Hazard ratios and 95% confidence intervals shown are adjusted for antihypertensive treatment, established risk factors (as defined in Table 3), and office blood pressure (model 4) or adjusted for antihypertensive treatment, established risk factors, and 24-hour ambulatory blood pressure (model 5). P.05 was considered statistically significant. P.05. Without myocardial infarction before baseline or during follow-up. Defined as night-day ambulatory blood pressure ratio of at least 1. 2006 American Medical Association. All rights reserved. (Reprinted) JAMA, June 28, 2006 Vol 295, No. 24 2865

prevalent CHF at baseline was defined as previous hospitalization for CHF. Since nondipping is associated with prevalent CHF, it is plausible that part of the association between nondipping and incident CHF identified in this study could be a result of undiagnosed prevalent CHF at baseline that only later was severe enough to require hospitalization. However, in a substudy of 343 participants without prevalent CHF, only 5 (1.5%) had an ejection fraction less than 0.40 on echocardiography; only 1 of these 5 had nondipping BP, and 3 developed CHF. Therefore, even if CHF were not diagnosed, it would have been uncommon. CONCLUSION In conclusion, a nondipping nightday BP pattern and an increased nighttime BP predicted CHF incidence independent of established risk factors in our large, community-based sample of elderly men. Nondipping was also a risk factor for CHF even when taking conventional office BP measurement into account. Nighttime BP appears to convey additive risk information about CHF, but its clinical value remains to be established in future studies. Author Contributions: Dr Ingelsson had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Ingelsson, Björklund- Bodegård, Lind, Ärnlöv, Sundström. Acquisition of data: Ingelsson, Björklund-Bodegård. Analysis and interpretation of data: Ingelsson, Björklund-Bodegård, Lind, Ärnlöv, Sundström. Drafting of the manuscript: Ingelsson, Björklund- Bodegård. Critical revision of the manuscript for important intellectual content: Björklund-Bodegård, Lind, Ärnlöv, Sundström. Statistical analysis: Ingelsson, Björklund-Bodegård, Ärnlöv, Sundström. Obtained funding: Ingelsson. Administrative, technical, or material support: Björklund-Bodegård. Study supervision: Lind, Sundström. Financial Disclosures: Dr Lind is a part-time employee at AstraZeneca Research and Development, Mölndal, Sweden, and a part-time employee at Uppsala University (AstraZeneca has no interests in this project and has not provided any financial support). No other disclosures were reported. Funding/Support: Funding was provided by Primary Health Care in Uppsala County, the Swedish Heart Lung Foundation (Hjärt-Lungfonden), and the Thuréus Foundation. Role of the Sponsors: The funding sources had no role in the design and conduct of the study, in the collection, analysis, and interpretation of the data, or in the preparation, review, or approval of the manuscript. REFERENCES 1. McMurray JJ, Pfeffer MA. Heart failure. Lancet. 2005;365:1877-1889. 2. Berry C, Murdoch DR, McMurray JJ. Economics of chronic heart failure. Eur J Heart Fail. 2001;3:283-291. 3. Levy D, Kenchaiah S, Larson MG, et al. Longterm trends in the incidence of and survival with heart failure. N Engl J Med. 2002;347:1397-1402. 4. Levy D, Larson MG, Vasan RS, Kannel WB, Ho KK. The progression from hypertension to congestive heart failure. JAMA. 1996;275:1557-1562. 5. Staessen JA, Thijs L, Fagard R, et al; Systolic Hypertension in Europe Trial Investigator. Predicting cardiovascular risk using conventional vs ambulatory blood pressure in older patients with systolic hypertension. JAMA. 1999;282:539-546 s. 6. Verdecchia P, Reboldi G, Porcellati C, et al. Risk of cardiovascular disease in relation to achieved office and ambulatory blood pressure control in treated hypertensive subjects. J Am Coll Cardiol. 2002;39:878-885. 7. Clement DL, De Buyzere ML, De Bacquer DA, et al. Prognostic value of ambulatory blood-pressure recordings in patients with treated hypertension. N Engl J Med. 2003;348:2407-2415. 8. Björklund K, Lind L, Zethelius B, Berglund L, Lithell H. Prognostic significance of 24-h ambulatory blood pressure characteristics for cardiovascular morbidity in a population of elderly men. J Hypertens. 2004;22: 1691-1697. 9. Hedstrand H. A study of middle-aged men with particular reference to risk factors for cardiovascular disease. Ups J Med Sci Suppl. 1975;19:1-61. 10. Byberg L, Siegbahn A, Berglund L, McKeigue P, Reneland R, Lithell H. Plasminogen activator inhibitor-1 activity is independently related to both insulin sensitivity and serum triglycerides in 70-year-old men. Arterioscler Thromb Vasc Biol. 1998;18:258-264. 11. Appel LJ, Whelton PK, Seidler AJ, Patel AR, Klag MJ. The accuracy and precision of the Accutracker ambulatory blood pressure monitor. Am J Epidemiol. 1990;132:343-354. 12. O Brien E, Asmar R, Beilin L, et al. European Society of Hypertension recommendations for conventional, ambulatory and home blood pressure measurement. J Hypertens. 2003;21:821-848. 13. Pickering TG, Hall JE, Appel LJ, et al. Recommendations for blood pressure measurement in humans and experimental animals, I: blood pressure measurement in humans: a statement for professionals from the Subcommittee of Professional and Public Education of the American Heart Association Council on High Blood Pressure Research. Circulation. 2005;111:697-716. 14. Staessen JA, Bieniaszewski L, O Brien E, et al; Ad Hoc Working Group. Nocturnal blood pressure fall on ambulatory monitoring in a large international database. Hypertension. 1997;29:30-39. 15. Ohkubo T, Hozawa A, Yamaguchi J, et al. Prognostic significance of the nocturnal decline in blood pressure in individuals with and without high 24-h blood pressure: the Ohasama study. J Hypertens. 2002; 20:2183-2189. 16. Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care. 2003;26(suppl 1):S5- S20. 17. Prineas RJ, Crow RS, Blackburn H. The Minnesota Code Manual of Electrocardiographic Findings: Standards and Procedures for Measurement and Classification. Bristol, England: John Wright; 1982. 18. Lindblad U, Råstam L, Ranstam J, Peterson M. Validity of register data on acute myocardial infarction and acute stroke: the Skaraborg Hypertension Project. Scand J Soc Med. 1993;21:3-9. 19. Task Force on Heart Failure of the European Society of Cardiology. Guidelines for the diagnosis of heart failure. Eur Heart J. 1995;16:741-751. 20. Ingelsson E, Ärnlöv J, Sundström J, Lind L. The validity of a diagnosis of heart failure in a hospital discharge register. Eur J Heart Fail. 2005;7:787-791. 21. Verdecchia P, Porcellati C, Schillaci G, et al. Ambulatory blood pressure: an independent predictor of prognosis in essential hypertension. Hypertension. 1994;24:793-801. 22. Ohkubo T, Imai Y, Tsuji I, et al. Relation between nocturnal decline in blood pressure and mortality: the Ohasama Study. Am J Hypertens. 1997;10: 1201-1207. 23. Chen YT, Vaccarino V, Williams CS, Butler J, Berkman LF, Krumholz HM. Risk factors for heart failure in the elderly: a prospective community-based study. Am J Med. 1999;106:605-612. 24. Gottdiener JS, Arnold AM, Aurigemma GP, et al. Predictors of congestive heart failure in the elderly: the Cardiovascular Health Study. J Am Coll Cardiol. 2000; 35:1628-1637. 25. Goyal D, Macfadyen RJ, Watson RD, Lip GY. Ambulatory blood pressure monitoring in heart failure: a systematic review. Eur J Heart Fail. 2005;7:149-156. 26. Verdecchia P, Schillaci G, Guerrieri M, et al. Prevalence and determinants of left ventricular diastolic filling abnormalities in an unselected hypertensive population. Eur Heart J. 1990;11:679-691. 27. Galderisi M, Petrocelli A, Alfieri A, Garofalo M, de Divitiis O. Impact of ambulatory blood pressure on left ventricular diastolic dysfunction in uncomplicated arterial systemic hypertension. Am J Cardiol. 1996;77:597-601. 28. Andersen NH, Poulsen SH, Poulsen PL, et al. Left ventricular dysfunction in hypertensive patients with type 2 diabetes mellitus. Diabet Med. 2005;22:1218-1225. 29. Higashi Y, Nakagawa K, Kimura M, et al. Circadian variation of blood pressure and endothelial function in patients with essential hypertension: a comparison of dippers and non-dippers. J Am Coll Cardiol. 2002;40:2039-2043. 30. von Känel R, Jain S, Mills PJ, et al. Relation of nocturnal blood pressure dipping to cellular adhesion, inflammation and hemostasis. J Hypertens. 2004;22: 2087-2093. 31. Lee KW, Blann AD, Lip GY. High pulse pressure and nondipping circadian blood pressure in patients with coronary artery disease: relationship to thrombogenesis and endothelial damage/dysfunction. Am J Hypertens. 2005;18:104-115. 32. Chong AY, Blann AD, Patel J, Freestone B, Hughes E, Lip GY. Endothelial dysfunction and damage in congestive heart failure: relation of flow-mediated dilation to circulating endothelial cells, plasma indexes of endothelial damage, and brain natriuretic peptide. Circulation. 2004;110:1794-1798. 33. Tousoulis D, Charakida M, Stefanadis C. Inflammation and endothelial dysfunction as therapeutic targets in patients with heart failure. Int J Cardiol. 2005; 100:347-353. 34. Sica DA. What are the influences of salt, potassium, the sympathetic nervous system, and the reninangiotensin system on the circadian variation in blood pressure? Blood Press Monit. 1999;4(suppl 2):S9-S16. 35. Sherwood A, Steffen PR, Blumenthal JA, Kuhn C, Hinderliter AL. Nighttime blood pressure dipping: the role of the sympathetic nervous system. Am J Hypertens. 2002;15:111-118. 36. Bell DS. Heart failure: the frequent, forgotten, and often fatal complication of diabetes. Diabetes Care. 2003;26:2433-2441. 37. Javaheri S, Parker TJ, Liming JD, et al. Sleep apnea in 81 ambulatory male patients with stable heart failure: types and their prevalences, consequences, and presentations. Circulation. 1998;97:2154-2159. 38. Bradley TD, Logan AG, Kimoff RJ, et al. Continuous positive airway pressure for central sleep apnea and heart failure. N Engl J Med. 2005;353:2025-2033. 39. Ziegler MG. Sleep disorders and the failure to lower nocturnal blood pressure. Curr Opin Nephrol Hypertens. 2003;12:97-102. 2866 JAMA, June 28, 2006 Vol 295, No. 24 (Reprinted) 2006 American Medical Association. All rights reserved.