JACC Vol. 22. No. 4 (SupplementA) 6A October 1!893:6A43A

PART II: NEW INSIGHTS INTO THE EPIDEMIOLOGY AND PATHOPHYSIOLOGY OF FAILURE BertramPitt, MD, FACC, Chairrnaro

EPIDEMIOLOGY

The Epidemiology of : The Framingham Study

KALON K. L. HO, MD,*? JOAN L. PINSKY, MA,* WILLIAM B. KANNEL, MD, FACC,g DANIEL LEVY, MD, FACC*tt Boston and Framingham, Massachusettsand Bethesda, Maryland

-

Congestive heart failure has become an incre&@ frequent women, whereas the age-adjustedprevoience of overt heart hilure reasonforhospiQla~durin~thc~2decpdosandderrly was 24/1,000in men and 25/1,000in women. Ikpite improved npresentsa~rlM!IdthplWblem.Dat?lfWltlWFkalUi@am tr&Wnts ror ischemil! heart d&ease and ,the HeartStudy~tbattbeinddenceofamgestiveheartfailure age-adjustedincidence of heart failure has declined by only increaseswithageandisbigherinmenthaninwomen.Hyper- ll%/calendar decade in men and by 17Walendar decade in temionandcoromuyhemtdkasearethetwomostcommon women during a 40-year period of observation.In addition, conditionspndrrt& its onset. Diabetesmellitus and eleckowr- eve heart failure rem&s highly lethal, with a median diograpldcleftventrkukhypcWophyarealsnassociatedwitb s~~~eaf1.7yeus~menend3.2y~inwomenanda anhnzea&riskofkartfailm~.Duringthe1!J8Os,theanmud 5-yearsurvival rate of 25% in men and 38% in women. ~ustedincidenceofcongestiveheartfailureamongpersons (J Am CoUCar&l 1993$2~trpplemeatA].&&13A) aged 245 years was 7.2 cases/l,000in men and 4.7 cases/l,000in

Although there is no consensus on the criteria required to (4-8) showed a steady increase in the number of hospital establish a diagnosis of congestive heart failure, it is clear admissions for congestive heart faihue during the last 2 that congestive heart failure represents a major health prob- decades. For example, the rate of hospital discharges with a lem. It has been estimated that congestive heart failure principal diagnosis of congestive heart faihue among Amer- afilicts nearly 4 million Americans, with soO,ooOnew cases icans 265 years increased from 7.5/1,0 in I%8 to l&3/ each year (I). Congestive heart failure was listed as the 1,Oo in 1989(4,7). principal cause for 37,400 deaths in 1988and is thought to In 1989,643,OOtlhospital discharges (2%of all discharges) have been a contributing cause of another 200,OtMdeaths carried a principal diagnosis of congestive heart failure (7. (2,3). Data from the National Hospital Discharge Survey An additional 4% (1.2 million) of all discharges carried a secondary diinosis of congestive heart failure. Amon; patients 265 years, congestive heart faihue was the principal From the *CharlesA. Dana Research Institute and the Harvard- diagnosis in 5% of those discharged and a secondary diag- ThomdiheLaboratory of the Departmentof ,Cardiovascular Divi- nosis in another 10%. The rate of hospital stays with any sion, Beth IsraelHospital and HarvardMedical School, Boston,Massachu- discharge diagnosis of congestive heart faihue increases setts:t’fhe Ftamh&amHeart Study of the NationalHeart, Lung, and Blood Institute,Framingham, ; #Division of Epidemiologyand Clin- dramatically with age, rising from 6.2/l ,008 in persons aged ical Applications,National Heart, Lung, and Blood Institnte, National 45 to 64 years to 47.411,OOtlin persons 265 years of age in Institutesof Health,Bethesda, Maryland and !lDepwtmentof Epidemiology 1989.Women were mnre frequently admitted to the hospital andReventlve Medicine,University Hospital and School for congestive heart fake than were men. In 1989,the rate of Medicine,Boston. Massachusetts. This study was suppotted by Contract NOI-HC-38038and Grant ST32 HL 0737413from the NatlonalInstitutes of of hospital admissions with any discharge diagnosis of COD Health,Bethesda, Maryland. gestive heart failure was 6.811,Oogin men and 8/l,ooO in Manuscriptreceived September 13, 1992;revised manuscriptreceived February10.1993, accepted March 22,1993. women; the rate of hospital admission with a principal for W Daniel Levy, MD, FraminghamHeart diagnosis of congestive heart faihue was 2.5/1,00 in men Study,5 ThurberStreet, Framin8ham. Massachusetts 01701. and 2.7/1,000in women (7).

81993 by the American College of JACC Vol. 22, No. 4 (Supplement A) HO ET AL. 7A October 1993:6A-13A EPIDEMIOLOGY OF CONGESTIVE HEAAT FAILURE

Methods Table1. Ckeriafor Congestive Heart Failure*

In the Framingham Heart Study, we have had the oppor- Majorcriteria tunity to examine the incidence, prevalenceand prognosisof Paroxysmalnoctutnal dyspnea congestive heart failure during four decades of observation. Neck vein distension Our study group was composed of participants in the Rales Framingham Study and the Framingham Offspring Study. In Radiographic cardiomegaly @reasing heart size on chest X-ray film) 1948,5,209residents of Framingham, Massachusetts aged 28 Acute palmorary edema Third sound gallop to 62 years were enrolled in a prospective epidemiologic Increased central wxous pressure (>I6 cm water at the right atrium) study. The selection criteria and study design have been Ciitdation time 225 L reported elsewhere (9,lO). Members of this “original co- Hepatojugular retlax hort” have subsquently been evaluated at 2-year intervals Pulmonary edema, Gsceral congestion or cardiomegaly at autopsy with medical history, and laborator= Weighloss 24.5 kg in 5 days in respottse to treat- of CfiF tests. XII19?!, children of the original study participants and Minor Criteria spouses of these children, aged 6 to 70 years, were entered in Bilateral ankle edema the Framingham Offspring Study (11). Serial evaluations Nocturnal coug!~ were performed on members of the Framingham Offspring Dyspnea on ordinary exertion Study 8 and 12years after enrollment. The 5,209 members of Hepatomegaly the original cohort and the 5,135 members of the offspring Plearal elf&n cohort were eligible for inclusion in these analyszs. To avoid Decrease in vita) capacity by 33% from maximal value recorded Tachycardia (rate ~I20 beat&in) inclusion of overly high risk subjects in the offspring study, 881 members of the offspring cohort with only one natural The diagnosis of congestive heart failure (CKF) required that two major or one major and two mittor criteria be present concarrently. Minor aiterkt parent in the original cohort were excluded; such subjects were acceptable only if they could not be attriited to aaother medical were preferentially enrolled if that parent had a high risk condition. or premarure . Cases of congestive heart failure diagnosed at the tirst Framingham Heart Study examination were included in estimates of heart nosed if voltage criteria for left ventricularhypertrophy were failure prevalence but excluded from estimates of heart fulfilled in an ECG exhibiting lateral ST segment depression failure incidence. or T wave flattening or inversion (13). At each examination, interim cardiovascular disease Using information obtained at Fram&ham Heart Study events were identified by medical history, physical exami- examinations and from hospital and physician record:;. a nation, 12-lead electrocardiogram (ECG) and review of diagnosis of congestive heart faihm: was established by the medical records. Hospital and physician records were ob- simultaneous presence of at least tro major or one major tained for participants who did not appear for an examina- and two minor criteria for congestive heart failure (Table 1). tion. All possible cardiovascuhu events were reviewed by a Minor criteria were acceptable only if they couid not be committee of three physicians, using established Framing- attributed to another medical condition (12-14). The same ham Heart Study protocols and definitions (12,13).Coronary criteria for congestive heart failure were used throughout the heart disease events included pectoris, coronary N-year follow-up period for these analyses. insufficiency (unstable angina) and The incidence of congestive heart faihtre was estimated (12). A patient was considered to have preexisting valvular using the cross-sectional pooling method, with each 2-year heart disease if the examining physician noted a systolic interval between successive examinations treated as an murmur grade III/VI or louder, a diastolic murmur or a independent observation. Only individuals free of heart palpable thrill at any examination before the onset of heart failure were included in a Zyear observation pool. Subjects failure. Radiographic and echocardiographic data were not who failed to appear for an examination were assigned used in the diagnosis of coronary or valvular heart disease. “hypothetical” examination dates corresponding to their Elevated blood pressure was defined as a systolic blood biennial anniversary dates; only information concerning pressure 2160 mm Hg or a diastolic blood pressure interval cardiovascular disease events was attributed to 195 mm Hg on each of two successive determinations by a these hypothetical examinations. Ail of the observations physician during a clinic examination. Hypertension was were further classified by the calendar year of the examina- defined as the presence of either elevated blood pressure or tion and the age of the subject at the time of the visit. The ongoing pharmacologic treatment of a previously elevated relations between disease incidence and calendar yew, as blood pressure. Subjects with a fasting blood glucose level well as between dichotomous risk factor measures and 27.77 mmoyliter (140mg/dl) or a random blood glucose level calendar year, were estimated using gender-specilic logistic 85 1.10 mmol/liter (200 mg/dl) were considered .,I be dia- regression analyses stratifying by 5-year age groups. A betic, as were individuals using insulin or oral hypoglycemic summary regression coefficient over all ages was derived by agents. Definite ECG left ventricular hypertrophy was diag- averaging the agespecific coefficients, using as weights the 8A HO ET AL. JAW vol. 22, No. 4 6upplement A) EPIDEMIOLOGY OF CONGESTIVE HEART FMLURE October 1993:6A-13A

_ I,_- inverse of their variances. The standard error of this sum- n Men mary coefficient was calculated as the square root of the q Women inverse of the sum of the weights. Some extreme age groups (for example, age 40 or 290 years) were excluded from certain models because there were too few observations to permit analysis. Age-adjusted differences in the proportion of cases of congestive heart failure with various antecedent conditions were examined by gender, using Cochran- Mantel-Xaenszel statistics stratified by 5-year age groups. For analysis of survival, Cox proportional hazards regres- sion models were employed to investigate the relations among survival, gender, age and calendar year at the time of first diagnosis of congestive heart failure. Because the study group was aging during the study period, in those models 39-39 4043 m-59 94369 70-79 so-89 investigating the effect of calendar year, analyses were stratified by age at diagnosis (in S-year age groups); simi- Age (w-) larly, in those models evaluating age at diagnosis, analyses were stratilied by calendar year, expressed in half-decades. Figure1. Incidencerates of congestiveheart failure (CHF) among FraminghamHeart Study subjects,by genderand age. Except for the model exploring the effect of gender on survival, men and women were analyzed separately. Median survival times and survival rates at 90 days and at 1,2,4,5 adjusted annual incidence of congestive heart failure was and 10 years after the diagnosis of congestive heart failure 2.3 cases/l,000 in men. In women, the corresponding age- were estimated using Kaplan-Meier methods. To facilitate adjusted annual incidence was 1.4 cases/l ,000. Among indi- comparison with results from heart failure intervention tri- viduals aged ~45 years, the age-adjusted annual incidence of als, we repeated the survival analyses excluding subjects congestive heart failure was 7.2 cases/l,000 in men and who died or were lost to follow-up during the 1st 90 days 4.7 eases/1,000in women. Analyses using logistic regression after the diagnosis of heart failure. Except where indicated, models revealed that the age-adjusted incidence of conges- age-adjusted rates were obtained by direct adjustment to the tive heart failure in men decreased from 1948 to 1988 by United States 1991population (15). The software package 1l%/calendar decade (p < 0.05); in women, the age-adjusted S-PLUS (Statistical Sciences, Inc.) was used for the survival incidence of congestive heart failure decreased by 17%/ analyses; all other computations were performed utilizing calendar decade (p c 0.05) (16). the SAS System (SAS Institute Inc.). A two-sided probabil- Prevalenceof heart fGlure. The prevalence of congesk ity (p) value 5 0.05 was required for statisticalsignificance. heart failure also increased with age (Fig. 2). Among men, Numeric values are expressed as mean value + SD. the prevalence of heart failure climbed from 8 cases/l ,000 in those 50 to 59 years of age to 66 cases/l ,000in those aged 80 ReSUltS Conges& heart lWlure ~poe&Among 9,405 Fmmingham Figure2. Prevalencerates of congestiveheart failure (CHF) among Heart Study participants(47% male) followed up from Septem- FraminghamHeart Study subjects,by genderand age. ber 1948to June 1988,congestive heart failure developed in 652. The mean age at the diagnosis of heart failure was 70.0 f 10.8years. These 331 men and 321 women were followed up Men for a median of 1.8 years after the diagnosisof congestive heart Women fails (mean 3.9 + 5.4 years; range 0 days to 35.8 years). There were 17 additional cases of congestive heart fake diagnosed at the lkst Framingham Heart Study examination. Incidence of cungestlve heart failure. The incidence of congestive heart failure increased dramaticallywith age (Fig. 1). The annual incidence increased from 3 cases/l ,000 in men aged 50 to 59 years to 27 cases/l ,000in men aged 80 to 89. In women, the annual incidence increased from 2 cases/l ,000in those 50 to 59 years of age to 22 cases/l,000 in those 80 to 89 years of age. The incidence of congestive heart failure was one-third lower in women than in men tier adjustment for 1 age (age-adjusted odds ratio of women/men 0.6; 95% co& 39ci9 49.49 59-59 69-99 m-79 go-89 dence interval [CII 0.5 to 0.7). During the 198Os,the age- Age (years) JACC Vol. 22. No. 4 (Supplement A) HO ET AL. 9A October 1993:6A-13A EPIDEMIOLOGY OF CONGESTIVE HEART FAILURE

0 Risk Factor Absent MEN WOMEN Fln 0.9

‘1. 1.0

‘4. 1.9

‘4. 2.0’

‘14. t.9

50 40 30 20 10 10 20 30 40 II = 331 q Hypertension alone n = 321 Ap-AdjusledIncidence al CHF. Age-Adjusied Incidence01 1 IF. per lODo pmmyear* pn low persat-year$ wCHD+HTN q Comnary Heart Dimwe alone Figure 3. Risk of congestive heart failure (CHF) in men in the FraminghamHeart Study by age and risk factor status. Relative Figure5. Prevalenceof coronaryheart disease (CHD) and hyper- risks(RR) for the developmentof heartfailure in the presenceof the tension(HTN) alone and in combinationamong Framingham Hean specified risk factor are displayed at the margins; values with Study subjectswith congestive heart failure, by gender. asterisksare significantat p < 0.0001.Cholesterol = serumcholes- terol >6.2 mmollliter(240 mg/dl); ECG-LVH = electrocardio- graphicleft ventricularhypertrophy. than those for younger men, but the excess risks were higher, reflecting greater absolute risk differences. In to 89 years. In women, the prevalence of heart failure women, neither an elevated serum level nor increased finm 8 cases/l,000 in those aged 50 to 59 years to cigarette resulted in an increased risk of heart 79 cases/l,000 in those 80 to 89 years of age. During the failure. In younger women, hypertension was associated 198Os,the age-adjusted prevalence of congestive heart fail- with a 3-fold increase in the incidence of heart failure, ure was 7.4/1,000in men and 7.7/1,000in women. Among diabetes with an 8-fold increase and left ventricularhyper- individuals aged 145 years, the age-adjusted prevalence of trophy with a 13.fold increase. As in men, these three congestive heart failure was 2411,000(or 2.4%) in men and conditions were associated with higher excess risks but 25/i ,000 (2.5%)in women. lower relative risks in older compared with younger women. Risk factors for heart failure. Figures 3 and 4 demon- Left ventricular hypertrophy was associated with an in- strate the risk of several conditions on the development of creased incidence of heart failure, even after controlling for congestive heart failure. In men, a serum cholesterol level blood pressure. Similarly, the effect of diabetes was inde- >6.2 mmol/liter (240 mgldl) was not associated with an pendent of concomitant hypertension or coronary heart increased risk of heart failure. Cigarette smoking increased disease. Other findings associated with an increased inci- the likeiihood of heart failure in younger, but not in older dence of overt congestive heart failure included , a men. In younger men, hypertension and diabetes mellitus high ratio of total cholesterol to high density lipoprotein were associated with a 4-fold increase in the incidence of cholesterol, proteinuria and ECG intraventricular conduc- heart failure and ECG left ventricularhypertrophy with a tion disturbances or nonspecific repolarization abnormali- IS-fold increase. In men older than 65 years of age, the ties. relative risks for heart failure associated with hypertension, Among the 331 men and 321 women in the Framingham diabetes and ECG left ventricular hypertrophy were less Heart Study who developed congestive heart failure during the follow-up period, hypertension and coronary heart dis- ease were the two most common preexisting conditions (Fig. Figure4. Riskof congestiveheart failure (CHF) in women in the FraminghamHeart Study by age and risk factor status. Relative 5). Seventy percent of men and 78% of women with heart risks(RR) for the developmentof heartfailure in the presenceof the failure had an antecedent diagnosis of hypertension. Forty specified risk factor are displayed at the m-s; values with percent of men and women with heart failure had a prior asterisksare significantat p < 0.0001.Definitions as in Fiiure 3. history of both hypertension and coronary heart disease. Prevalent coronary heart disease was less common in IIIRisk F-Absent W RiikFactorP womenthan in men, with an odds ratioof 0.S after adjusting RR RR es64 for age (95%CI 0.40 to 0.76); it was found in 59%of men and 0.7 0.8 48% of women with new congestive heart failure. Eleven 1.1 1.3 percent of men and 15%of women with heartfailure had no

-3.0 1.9’ prior history of hypertension or coronary heart disease. The prevalence of several causes of and risk factors for 7.7 3.6’ heart failure are shown in Fire 6. Again, hypertension and

‘128 5.4’ coronaryheart disease were the two most prevalent co& ~ 50 40 30 20 10 0 10 20 30 40 50 tions in subjects with new heart failure. Recent cigarette reb*I5lmd hcidma of CHF, Ago-AdjuM&dlneidan~a of CHF. pm 1wD pers-sn pa 1oDD plsoycam smoking was less common in women than in men, with an JACC Vol. 22. No. 4 (Supplement A) IOA HO ET AL. EPIDEMIOLOGY OF CONGESTIVE HEART FAlLVRE October 1993:6A-13A

!hvivRl after diagnosisof heart failure, Among the 652 persons with congestive heart failure, 551 deaths (84.5%) occurred. The median survival time after the diagnosis of congestive heart failure was 1.66 years in men and 3.17 years in women (Table 2) (17).Overall l-year survival rates in men and women were 57% and 64%, respectively. The overall S-year survival rate was 25% in men and 38% in women. There was no significanttemporal change in overall survival after the diagnosis of congestive heart failure during the &year period of observation (age-adjusted male hazards ratio for mortality = l.O8/calendardecade, 95% CI 0.92 to 1.27; age-adjusted female hamds ratio for mortality = 40 so so 100 1.02lcalendardecade, 95% CI 0.83 to 1.26). After adjusting Pmvdence (36) for age, survival after the development of heart failure was Flgan 6. Prevalenceof certain conditions among Framingham better in women than in men, with a hazards ratio of 0.64 Heart Study subjects with congestive heart failure, by gender. (95% CI 0.54 to 0.77). The mortality rate increased with ECG-LVH= electrocardiographicleft ventricularhypertrophy. advancing age at the diagnosis of heart failure. In men, the increase in mortality rate was 27%/decadeof age (95%CI 9% to 47%). In women, the mortality rate increased by 61%/ odds ratio of 0.51 after adjusting for age (95%CI 0.32 to 0.79); decade of age (95%confidence interval 37% to 90%). overall, 42% of men and 24% of women smoked cigarettes at the time of the Framingham Heart Study examination imme- diately before the diagnosis of heart failure. Electrocardio- Discussion graphic left ventricular hypertrophy was noted before the Data from the Framingham Heart Study indicate that diagnosis of heart failure in 17%of men and 18%of women. congestive heart failure is more commons in older persons After adjustingfor age, only coronary heart disease and recent and affects approximately 2.5% of the population aged 245 cigarette smoking were signihcantly less common in women years. Hypertension preceded the onset of heart failure in 70% than in men with new congestive heart failure. of men and 78% of women and was associated with a two- to From 1948to 1988,the age-adjusted prevalence of coro- four-fold increase in the incidence of heart failure. Coronary nary heart disease among men with new congestive heart heart disease was present in 59% of men and 48% of women failure increased by 46%/calendar decade (p < 0.05). In with heart failure and has been increasingin prevalence among contrast, the age-adjusted prevalence of coronary heart new cases of heart failure, Diabetes mehitus and ECG lefi disease among all men in the Framingham Heart Study ventricular hypertrophy were also associated with a signiti- decreased by 8%/calendar decade (p < 0.05). The results cantly increased incidence of overt congestive heart failure. were similar in women, who had an age-adjusted46%/ Despite tie increased use of improved therapies for ischemic calendardecade increase in the prevalence of coronary heart heart disease and hypertension, the age-adjustedincidence of disease among subjecrs with new heart failure (p < 0.05) in heart failure has declined by only 1l%/calendar decade in men association with an age-adjusted 16blcalendar decade de- and by 17% per calendar decade in women during a #year cline in the prevalence of coronary heart disease among all period of observation. In addition, congestive heart failure women in the Framingham Heart Study (p < 0.001). remains highly lethal, with a median survival time of 1.7 years

Tabte2. OverallSurvival AtIer Congestke HeartFailure as Estimatedby Kaplan-MeierMethods

M&Ill Survival Survival Rates Subjects Time (no.) (yr) 90 days 1 Yr 2yr 5 Yr IO yr AU subjects with congestive hean failure Men 331 1.66 0.73 0.57 0.46 0.25 0.11 Women 321 3.17 0.72 0.64 0.56 0.38 0.21 Subjects with congestive heart failurewho surv:ved~90 days MCll 237 3.21 e- 0.79 0.63 0.35 0.15 Women 230 5.39 - 0.88 0.78 0.53 0.29 JACC Vol. 22, No. 4 (Supplement A) HO ET AL. 1lA Occtcvb,ber19B:GA-13A EPIDEMIOLOGY OF CONGESTIVE HEART FAILURE

Table3. EpidemiologicStudies of CongestiveHeart Failure

Comparable Anuual Compamble CHF Annual CHF CHF lncidcilce in CHF Prevalence in htcidence Framingham Prevalence Framingham Study Group Age tyrt (per 1.000~ lper l,OOM (per LOOM (per 1,009) Tecumseh, Michigan: 1959-1960 (32) Men 40-79 24= 15* Women 40-79 36* 13* Evans County, Georgia: 1960-1962 (31) Men 4f+74 28* 16’ Women 45-74 18* is* Rural Vermont and North Carolina: 1962-1963 (27) Men and women All ages I .4* I.42 Men All ages IO' 8* Women All ages II* 6* CZteborg, Sweden: l%3-l9gO (28,291 Men 50-54 1.5 1.8 55-60 4.3 4.5 61-67 10.2 6.4 50 21 7 54 24 9 66 43 I6 67 130 32 CMteborg, Sweden: 1971-1972 (33) Men 70 110 48 75 170 50 Women 70 80 35 75 110 53 NHANES-I: 1971-1975 (34) Men 25-54 8 I 5>64 45 12 6W4 48 39 Women 25-54 13 1 55-64 30 14 65-74 43 31 Framingham: 1980-1988 Men All ages 2.3* 7* 845 7.2* 24* Women All ages I .4* 8* 245 4.7* 25* Rochester. Minnesota: 1986 (30) Men 235 12* 16* Women 235 11* 16* Eastern Finland: 1986-1988 (26) Men 45-74 4.0-4.2* 4.6* Women 45-74 l.O-l.6* 2.2*

*Age-adjusted to the United States 1991 population. Numbers in parentheses indicate reference numbers. CHF = congestive heart failure; NHANRS-I = Fist National Health and Nutrition Examination Survey. in men and 3.2 years in women. With 5-year survival rates of series utilize selected patients who survived the acute onset of 25%in men and 38%in women, heart failure is associEiedwith congestive heart failure to achieve study entry. a shorter life expectancy than that of many common malignan- Other epidemiokygicstudies of heartfailure. There is no cies. agreement on the precise criteria required to establish a The Framingham Heart Study provides several unique diagnosis of congestive heart failure. Most prior heart failure advantages for the study of the epidemiology of congestive intervention trials (18-22) required a reduced left ventricular heart failure. A large unselected patient population with nearly ejection fraction as an entry criterion; therefore, their results equal numbers ofmen and women was observed for >40 years may be applicableonly to patients with systolic heart failure. with uniform ascertainment of events. Ah persons with con- Congestive heart failure has always been defined in the gestive heart faikn were followed up from the time of diagno- Framingham Heart Study on the basis of clinical criteria that sis; in contrast, intervention trials and hospital-based referral have been previously validated (23)and are compatible with JACCVol. 22, No. 4 (SupplementA) 12A HOET AL. EPBJElUIOLOGYOF CONGESTIVEHEART ‘FAILURE October1993M-13A the presence of either diastolic or systolic dysfunction sons who would have been excluded from many of the recent (24,22). Gther epidemiologic studies of heart failure have pharmacologic heart failure trials (for example, those pa- used different definitions of heart failure; nonetheless, some tients with acute pulmonary edema, unstable angina, recent oftheir resultsare similarto those from Framingham (Table myocardial infarctions or significant valvular heart disease). 3). For example, from 1986to 1988in eastern Finland, the When the analysis was restricted to subjects who survived age-adjustedannual incidence of heart failure among persons ~90 days after the diagnosis of heart failure, a population 45 to 74 years old was 4 to 4.2 cases/l,000 in men and 1.O to more comparable to those included in intervention trials, the 1.6 cases/l,ooOin women using either the Framingbam or median survival time was 3.21 years in men and 5.39 years in modiied Boston criteria for heart failure (26).The contp- women. In men, the l-year survival rate was 7% and the ble annual incidence of congestive heart failure in Framing- 5-year survival rate was 35%. In women, the l- and S-year ham during the 1980swas 4.6 cases/l,000 in men and 2.2 survival rates were 88% and 53%, respectively. These re- cases/l,000 in women. Using a simple survey, Gibson et al. sults more closely resemble those from heart failure inter- (27) found that the age-adjusted annual incidence of heart vention trials. For example, the l-, 2- and 4-year survival failure in two rural counties in 1962 and 1963 was 1.4 rates for men in the placebo group in the first Veterans cases/l,000 (adjusted to the U. S. 1991 population); the Administration Cooperative Vasodilator-Heart Failure Trial comparably age-adjusted annual incidence of congestive (V-HeFT I) were 81%, 66% and 46910,respectively (19). In heart failure in Framingham in the 1960s was also 1.4 our study, the corresponding survival rates were 79%, 63% cases/l,O. The Swedish study (2829) of male residents of ard 45% in men who survived ~90 days after the diagnosis Giiteborg born in 1913 also reported incidence rates for of heart failure. congestive heart failure (using a heart failure score) that are Concluslons. Despite advances in our therapies for isch- very similar to those from Framingham. The prevalence emic heart disease and hypertension, congestive heart fail- rates of heart failure in epidemiologic studies are more ure remains a common and highly lethal condition. There is disparate. The prevalence of heart failure among residents of accumulating evidence that treatment of hypertension and Rochester, Minnesota 235 years of age in 1986was 1.2% in left ventricular systolic dysfunction can decrease the inci- men and 1.1%in women (30); the comparable prevalence of dence of clinical congestive heart failure (22,35-37)and that congestive heart failure in Framinghatr in the 1980s was use of angiotensin-converting enzyme inhibitors or vasodi- 1.6%in both men and women. In Evans County, Georgia in lators can prolong survival after the onset of heart failure 1960to 1962, heart failure was found in 2.8% of men and (19~21,38).Clearly, the diagnosis of congestive heart failure 1.8% of women aged 45 to 74 years (31); the comparable encompasses a constellation of symptoms that can result age-srdjusted prevalence in Framingham in the 1960s was fim a variety of pathophysiologic processes retlectiug both 1.6% in men and 1.5% in women. In Tecumseh, Michigan systolic and diastolic dysfunction. Improved strategies for among persons aged 40 to 79 years in 1959and 1960, the the prevention and treatment of congestive heart failure age-adjusted prevalence of heart failure was 2.4%in men and must take into account the heterogeneous nature of this 3.6% in women (32); the respective rates in Framingham clinical syndrome. were 1.5% and 1.3%. Several Swedish studies (28,29,33) have used more liberal definitions of heart failure and consequently documented prevalence rates of heart failure References up to four times higher than those in Framingham. 1. MassieBY, PackerM. Congestiveheart failure:current controversies Over a #year period of observation, the age-adjusted andfuture prospects. Am J Cardiol199o,ti429-30. 2. NationalCenter for HealthStatistics. Vital Statistics of the United States, incidence of congestive heart failure decreased by ll%/ 1988.Vol Il. Mwtality,Fart A. Washington,DC: Public Health Service. calendar decade in men and by 17%/calendardecade in 1991;DHHS pubhcatioa no. (PHS)91-1101. women. Analyses are ongoing to further elucidate the rela- 4. GiIIumRF. Heartfaiktre in the UnitedStates 1970-1985. Am Heart J tions between the incidence and prevalence of congestive 1987;113:1043-5. 4. NationalCenter for Heahh Statistics. Inpatient Utilization of Short-Stay heart failure and temporal changes in the prevalence of HospitaIsby Diagwsis:United States, 1968-1980.Rockviie andHyatta coronary heart disease, valvular heart disease, hyperten- vilIe.Marykut~ NatIooal Center for HealthStatistics, 1974-1984, DHEW sion, diabetes and other conditions. andDHHS pubIicatkms (Vital and HoaltbStatIstIcs; series 13). 5. Utilizationof short-stayhospItaIs. United States, 1979-1986,Annual !!Jur&al afterdlagwsk uf heart falhtre. We were unable Summary.Hyattsviie. Maryland:National Center for HealthStatistics, to detect any significant improvement in survivalafter the IWI-1988;DHHS publications (Vital and HealthStatistics: series 13). diagnosis of congestive heart failure during the rlO_year 6. NationalCenter for Health Statistics. Detailed Diagnoses and Prow Period of observation. However, our observational study did dures, National Hospital DischargeSurvey, 1983-1988.Hyattsville. Maryland National Center for Health Statistics, 1985-1991;DHHS not directly examine the impact of therapy on survival after publications(Vital and HealtbStatistics; series 13). the onset of heart failure. In addition, most of the follow-up 7. GravesEJ. DetailedDiioses and Procedures,Natioaal wtal Dis- period for our study occurred before the widespread use of chargeSurvey, 1989.Hyattsville, Maryland: National Center far Heakb Statistics,1991; DHHS publication no. (PHS)91-1769, (Vital and Health VaSodi~tOrS,angiotensin-converting enzyme inhibitors and Statistics;series 13;no. 108). heart transplantation. Gf note, our populationincluded per- 8. GhaliJR, CooperR, Ford E. Trendsin hospitalizationrates for heart JACC vol. 22, No. 4 (Supplement A! HO ET AL. 13A October 199J:fiA-13A EPIDEMIOLOGY OF CONGESTIVE HEART FAILURE

failure in the United States, 1973-1986: evidence for increasing population between lefl ventricular systolic bmction and congestive heart failure prevalence. Arch Intern Med 1990;150:769-73. diagnosed by clinical criteria. Circulation i!B&77$t?7-12. 9. Dawher TR. Meadors GF, Moore FE. Epidemiologic approaches to heart 24. Dougherty AH, Naccarelli GV, Gray EL, Hicks CH. Goldstein RA. disease: the Framingham study. Am J Public Health 1951;41:279-86. Congestive heatt failure with normai systolic function. Am J Catdiol 10. Dawber TR. Kannel WB, Lye8 LP. An approach to longitudinal studies l984;54:778-82. in a community. the Frami&am study. Ann NY Acad Sci 1963;107:539- 25. Grossman W. Diasto!ic dysfunction in congestive heart failure: me&a- 56. nisms and clinical implications. N Engl J Med 1991;325:1557-64. Il. Kannel WB. Feinleib M. McNamara PM, Garrison RJ, Castelli WP. An 26. Remes J. Reunanen A, Aromaa A, Py6r8lg K. incidence of heart failure investigation of coronary heart disease in families: the Framingham in eastern Fiinland: a population-based surveillance study. Eur Heart J Offspring Study. Am J Epidemiol 1979;110:281%J. 1992;13:588-93. 12. Shurtleff D. Some characteristics related to the incidence of cardiovas- cular disease and death: Framingham study M-year follow-up. In: Kannel 27. Gibson TC, White KL, Klainer LM. The prevalence of congestive heart WB, Gordon T, eds. The Framingham study: An Epidemiological Inves- failure in two rutal communities. J Chronic Dis 1%6,19:141-52. tigation of Cardiovascular Disease. Section 30. DHEW publication no. 28. Et&son H. Svgrdsudd K. Larsson B, et al. Risk factors for heart failure (NIH) 74-599. Washington DC: U.S. Government Printing Oflice, 1974: in the general population: the study of men bcm in 1913. Eur Heart J i7-40. 1989;10:647-56. 13. Cupples LA, D’Agostino RB. Some risk factors related to the annual 29. Wilhelmsen L, Eriksson H. Sv6rdsudd K, Caidahl K. Improving the incidence of cardiovascular disease and death using pooled repeated detection and diagnosis of congestive heart failure. Eur Heart J 198% biennial measurements: Framingham Heart Study, 30-year followup. In: Kgsuppl C): 138. Kannel WB, Wolf PA, Garrison RI, eds. The Framingham Study: An 30. Phillips SJ. Wbisnant JP, O’Faiion WM. Frye RL. Prevalence of cardio- Epidemiologicai Investigation of Cardiovascular Disease. Section 34. vascular disease and diabetes mellitus in residents of Rochester, Minne- DHHS publication no. (NIH) 87-2703. Washingtoa DC U.S. Government sota. Mayo Clin Proc 19!JO$5:344-59. Printing Office, 1987:9-20. 31. Garrison GE, M&no@ JR, Hames CG. Stulb SC. Prevalence of 14. McKee PA, Castelli WP, McNamara Phf. Kannel WB. The natural chronic congestive heart failure in the population of Evans County, history of congestive heart failure: the Framingham study. N Engl J Med Georgia. Am J Epidemiol 1%,83:338-44. 1971;285:1441-6. 32. Epstein FH, Ostrander LD, Johnson BC. et al. Epidemiological studies of IS. U.S. Bureau of the Census. Statistical Abstract of the United States: 1992 cardiovascular disease in a total community-Tecumseh, Michigan. Amt (112th ed). Washington DC: U.S. Government Printing Gfflce, 1992:15. Intcm Med 1965:6t:ll70-87. 16. Katmel WB, Pinsky J. Trends in cardiac failure-incidence and cause over three decades in the Framingham study (abstr). J Am Coil Cardiol 33. Landahl S, Svanborg A, Astrand K. Hearl volume and the prevalence of 1991;17387A. certain common cardiovascular disxders at 70 and 75 years of age. Eur Heart J 1984;5:326-3 I. 17. Ho KKL. Anderson KM, Kannel WB. Grossman W, Levy D. Survival after the onset of congestive heart failure in Framingham Heart Study 34. Schocken DD, Arrieta MI, Leaverton PE, Ross EA. Prevalence and subjects. Circulation 1993:88:107-15. mortality rate of congestive heart failure in the United States. J Am Coil 18. Gradman A, Deedwania P. Cody R. et al. Predictors of total mortality and Cardiol 1992:20:301-6. sudden death in mild to moderate heart failure J Am Coil Cardiol 35. Yusuf S, Thorn T. Abbott RD. Changes in hypertension treatment and in 1989;14:564-70. congestive heart failure mortality in the United States. Hypertension 19. Cohn JN. ArchlRxtldDG. Ziesche S, et al. Effect of vasodilator therapy on 1989;13(suppl I):I-74-9. mortality in chronic congestive heart failure: results of a Veterans 36. SHEP Cooperative Research Group. Prevention of stroke by antihyper- Administration Cooperative Study. N Engl J Med 1986,314:1547-52. iensive drug treatment in older persons with isolated systolic hyperten- 20. Cohn JN, Johnson G, Ziesche S, et. al. A comparison of enalapril with sion: final results of the Systolic Hypertension in the Elderly Program hydralazine-isosorbide dinitrate in the treatment of chronic congestive (SHEP). JAMA 1991;265:3255-64. heart failure. N Engl J Med 1991;325:303-IO. 37. Pfelfer MA, Braunwald E. Moye LA, et al. Elfect of captopril on 21. The SOLVD Investigators. Effect of enalapril on survival in patients with mortality and morbidity in patients with left ventricular dysfunction after reduced left ventricular ejection fractions and congestive heart my& infarction: results of the Survival and Ventricular Enlarge thilure. N Engi J Med 1991;325:293-302. meat Trial. N Engl J Med l!B2~27:669-77. 22. The SOLVD investigators. Effect of enalapril on mortality and the 38. The CONSENSUS Trial Study Group. Effects ofenalapril on mortality in development of heart failure in asymptomatic patients with reduced left severe congestive heart failure: results of the Cooperative North Scandi- ventricular ejection fractions. N Engl J Med 1992;327:68wl. navian Enalapril Survival Study (CONSENSUS). N Engi J Med 1987; 23. Marantz PR, Tobm JN, WassertheiiSmoller S, et al. The relationship 316:1429-35.