|
Author
Wilson S Colucci, MD
Section Editor
Stephen S Gottlieb, MD
Deputy Editor
Susan B Yeon, MD, JD, FACC
INTRODUCTION — Aging of the population and prolongation of the lives of cardiac patients by modern therapeutic innovations has led to an increasing incidence of heart failure (HF). Despite improvements in therapy, the mortality rate in patients with HF has remained unacceptably high [1]. (See "Epidemiology and causes of heart failure".)
The prognosis of patients with HF will be reviewed here. The many factors that can be used to predict survival and the prognosis in patients with asymptomatic left ventricular systolic or diastolic dysfunction are discussed separately. (See "Predictors of survival in heart failure due to systolic dysfunction" and "Evaluation and management of asymptomatic left ventricular systolic dysfunction" and "Treatment and prognosis of diastolic heart failure".)
STAGES OF HF — Heart failure is a progressive condition, beginning with predisposing factors and leading to the development and worsening of clinical illness. There are several stages in the evolution of HF, as outlined by an ACC/AHA task force [2]:
* Stage A — High risk for HF, without structural heart disease or symptoms
* Stage B — Heart disease with asymptomatic left ventricular dysfunction
* Stage C — Prior or current symptoms of HF
* Stage D — Refractory end stage HF
This staged system emphasizes the progressive nature of HF and defines the appropriate therapeutic approach for each stage (algorithm 1).
HOSPITALIZATION — The association of nonfatal hospitalization and subsequent mortality rates was studied using data on 7572 chronic HF patients with reduced or preserved LVEF in the CHARM trials [3]. Mortality rate was increased after HF hospitalizations, even after adjustment for baseline predictors of death (HR 3.2; 95% CI 2.8-3.5). The increased risk of death was highest within one month of discharge and declined progressively over time.
There is an appreciable readmission rate for decompensated HF and patients hospitalized longer or more frequently have a higher mortality rate [3-7]. In a review of elderly patients, for example, 8 percent required readmission for HF within six months of the initial hospitalization [8].
Poor compliance is an important contributing factor in many patients requiring readmission [4-6]. In one series, lack of adherence to the medical program (drug or diet) was the most common reason for readmission, occurring in 41 percent of cases; another 12 percent received inadequate preadmission treatment [4]. It has been estimated that more than one-half of readmissions are preventable [5].
IMPACT OF CARE PROVIDERS AND PROGRAMS — The long-term prognosis of patients with HF is improved by both subspecialist care and by nurse-directed disease management programs.
Effect of specialty care — Specialist care of HF patients by a cardiologist is more expensive than care provided by a generalist, due to the use of more diagnostic techniques and to longer hospitalizations. However, it has been hoped that this would be counterbalanced by an improvement in outcome.
Inpatient care — Observational studies have evaluated the impact of inpatient specialist care of patients with decompensated HF and some have found improved clinical outcomes [9,10]. The potential impact of specialty care was illustrated by a population-based study of 38,702 patients hospitalized with HF in Ontario, Canada between 1994 and 1996. Patients attended by cardiologists had lower1-year adjusted risk of death as well as composite risk of death or readmission compared to patients attended by noncardiologists [10].
However, the applicability of this data to contemporary practice is unclear. Since the studies were not randomized, and a number of confounding factors (eg, comorbidities and severity of illness) could have biased the results despite risk adjustment. Furthermore, the standard of care for HF patients has evolved significantly in recent years, leading both to more therapeutic options that impact long term outcomes (see 'Effect of medical therapy' below, and also to more complex decision making regarding the selection of an appropriate medical regimen.
Outpatient care — Patient outcomes appear to be improved when a cardiologist participates in outpatient HF management [11-13]. In a retrospective cohort study of 403 patients with new onset HF, cardiologist care (defined as at least two office visits or ≥25 percent office visits) was an independent predictor of a lower risk of the combined outcome of death or hospitalization at 24 months. Patients seen by a cardiologist were significantly more likely to have been treated with ACE inhibitors (83 versus 68 percent in those only seen by a primary care physicians) and beta blockers (38 versus 22 percent).
Disease management programs — HF disease management is a conceptual framework for the care of HF patients that typically involves some or all of the following:
* Discharge planning
* Multidisciplinary coordination
* Formal guidelines for patient management, based on trial data and clinical experience
* Formal patient education
* A program of frequent patient assessment and, for outpatients, frequent contact between the patient and care providers
These are usually nurse-directed, multidisciplinary programs. A number of studies and meta-analyses have demonstrated that these programs have the following significant benefits [2,14-30]:
* Improved quality of life
* Improved survival
* Reduced hospitalizations
Similar benefits have been shown in patients with far advanced disease (class III or IV HF) who have been accepted for heart transplantation (graph 1) [20].
One of the largest meta-analyses included 29 randomized trials and 5039 patients [26]. The following findings were noted:
* Strategies that incorporated specialized multidisciplinary follow-up reduced mortality (risk ratio [RR] 0.75), HF hospitalizations (RR 0.74), and all-cause hospitalizations (RR 0.81).
* Programs that emphasized patient self-care activities reduced HF hospitalizations (RR 0.66) and all-cause hospitalizations (RR 0.73) but had no effect on mortality.
* Studies that employed telephone contact reduced HF hospitalizations (RR 0.75) but not all-cause hospitalizations or mortality.
The long-term impact of these programs was assessed in a trial that included 297 patients who were randomly assigned to usual care or to nurse-led multidisciplinary home-based intervention (HBI) [22]. Over a minimum follow-up period of 7.5 years, HBI was associated with a significant increase in median survival (40 versus 22 months with usual care) and significant reductions in readmission rate (2.0 versus 3.7 admissions) and the number of hospital days (14.8 versus 28.4 days per patient-year). HBI was cost-effective, with a cost of only 1729 dollars per additional year of life gained. (See "A short primer on cost-effectiveness analysis".)
Adherence to guidelines — The impact of adherence to Joint Commission on Accreditation of Healthcare Organizations (JCAHO) HF core measures was evaluated in a study of 2958 patients in a 20-hospital system. A positive and incremental relationship between degree of adherence and one-year survival was demonstrated [31].
SURVIVAL IN HF — Morbidity and mortality after the onset of symptomatic HF is extremely high, although variable mortality rates have been reported [32-38], which in part reflect differences in disease severity and in appropriate medical therapy. The following statistics refer to patients with systolic HF, since the natural history of diastolic dysfunction is less well defined (see 'Diastolic dysfunction' below.
In-hospital mortality — Among patients hospitalized for HF, in-hospital mortality and length of hospital stay have decreased, despite an increase in the severity of HF [39-41]. As an example, a single-center study of 6676 patients hospitalized for HF found that, over a 10 year period (1986 to 1996), there was reduction in observed in-hospital mortality from 8.4 to 6.1 percent (despite slight increase in predicted mortality) and reduction in adjusted length of hospital stay from 7.7 to 5.6 days [39].
However, reductions in in-hospital mortality have been accompanied by lesser [40] or no decrease in 30-day mortality [41]. An analysis of 2,540,838 Medicare beneficiaries hospitalized with HF between 2001 and December 2005 revealed reduction in observed in-hospital mortality from 5.1 to 4.2 percent although 30 day mortality was unchanged at 11 percent. Nearly one in four patients was readmitted within 30 days of the index hospitalization.
The 30-day mortality data demonstrate that the use of in-hospital mortality alone in evaluating HF prognosis can be misleading, especially when the length of hospital stay is declining and readmission rates are high. They also raise the concern that reduction in the length of stay after hospitalization for HF has had adverse consequences, possibly increasing early post-discharge mortality because a greater number of patients are discharged in an unstable condition. However, it is also possible that in the later years of the studies more terminally ill patients were being discharged from the hospital to die in other settings.
Long-term mortality — Long-term mortality rates for patients with HF have improved over time [40,42,43], although later studies indicate a slower rate of improvement [41,44]. In the Framingham study, the following reductions in age-adjusted mortality from the time of onset of HF were noted from 1950 through 1969 compared to 1990 through 1999 [42]:
* The one-year mortality declined from 30 to 28 percent in men and from 28 to 24 percent in women.
* The five-year mortality declined from 70 to 59 percent in men and from 57 to 45 percent in women; the adjusted reduction in mortality was about 30 percent in both sexes.
* There was a significant overall trend of a 12 percent reduction in mortality per decade during this time period; almost all of the improvement in mortality occurred after 1980 and particularly after 1990.
A decline in HF mortality was also noted in an analysis from the Mayo Clinic [43]. For the periods 1979 to 1984 compared to 1996 to 2000, the one-year mortality fell from 30 to 21 percent in men and from 20 to 17 percent in women. The five-year mortality fell from 65 to 50 percent in men and from 51 to 46 percent in women. Survival improved most among younger men and least among older women.
Lower rates of improvement in HF mortality rates were observed in two later studies of Medicare beneficiaries spanning shorter periods of time [41,44]:
* In a study of 3,957,520 Medicare beneficiaries hospitalized with HF between 1992 and 1999, the observed one-year mortality rate decreased only slightly (from 32.5 to 31.7 percent) during the study period due to a decline in mortality between 1993 and 1994 [44].
* In a study of 2,540,838 Medicare beneficiaries hospitalized with HF between 2001 and 2005, the observed one-year mortality rate (37 percent) did not fall over time, although the adjusted hazard of mortality at one-year was slightly lower in 2005 than in 2001 (hazard ratio 0.98; 95% CI, 0.97-0.99) [41].
Mortality rates in the placebo arms of clinical drug trials, which represent a selected population, have been variable, depending in large part upon the severity of the HF. In the placebo arms of the angiotensin converting enzyme (ACE) inhibitor trials of patients with systolic dysfunction, in which few patients were taking the two other drugs known to improve survival, beta blockers and spironolactone, the mortality rate was 52 percent at one year for patients with New York Heart Association (NYHA) class III or IV disease [45] compared to 25 percent at two years for NYHA class II to III disease (table 1) [46,47]; among patients with asymptomatic left ventricular dysfunction, the incidence of cardiovascular death or symptomatic HF was 39 percent at just over three years (graph 2A-C) [48].
The entry criteria and selection biases explain the better outlook in these trials than in the population-based studies. Trial participants generally have fewer complicating illnesses, such as significant ischemic symptoms amenable to therapy or valve disease, fewer older patients, and are more likely to receive optimal therapy. As a result, trials are expected to overestimate survival.
Predictive models — A variety of predictors of survival have been identified in patients with HF, such as peak VO2, New York Heart Association (NYHA) functional class, left ventricular ejection fraction, and markers of the adequacy of tissue perfusion. (See "Predictors of survival in heart failure due to systolic dysfunction".)
Although these risk factors correlate with survival on a statistical basis in a large population, their ability to predict survival in individual patients is limited. As a result, a number of retrospective analyses have been used to develop predictive models that utilize multiple indicators to generate a more accurate estimate of prognosis [49-52].
Potential benefits of using prognostic models for HF include the following [53]:
* Enables patients and families to have a realistic expectation of the prognosis
* Enables appropriate allocation of resources, including transplantation, mechanical circulatory assist devices, and implantable defibrillators
* Enables selection of therapies most likely to positively affect the quality and quantity of life
* Promotes open, honest communication between clinicians, patients, and their families to define the goals of therapy.
Potential hazards of using prognostic models for HF include the following [53]:
* The model was derived from a different population of patients
* Patient compliance, preference, or attitudes are not incorporated
* New therapies become available, making the models obsolete
* The patient is not in compensated HF or is not on evidence-based therapies
* Uncertainty in applying the model to an individual patient cannot be quantified and this uncertainty may be difficult for clinicians to effectively explain to patients and their families
* Scores from the models replace informed, compassionate, clinician-patient conversations
Given the limitations of prognostic models, use of a prognostic model should supplement rather than replace the judgment of the clinical team [53]. A study of the predictive accuracy of physicians and HF nurses for estimating risk of hospitalization and death among patients with advanced HF found that nurse estimations of mortality added significantly to the derived prognostic model but physician estimations did not [54]. It was hypothesized that the nurses were better versed in patient psychosocial characteristics that impact HF outcomes.
EFFECT model — The EFFECT model was derived, tested, and intended to be used in patients hospitalized for HF [49]. The derivation cohort included 2624 patients in the EFFECT study, who presented with HF at 34 hospitals in Ontario, Canada between 1999 and 2001. The model was then validated in 1407 patients presenting between 1997 and 1999.
Multiple clinical characteristics, including both HF-related factors (respiratory rate, systolic pressure, blood urea nitrogen, and serum sodium concentration) and comorbidities (eg, COPD, anemia, malignancy) were correlated with 30-day and one-year mortality. Points were assigned to each significant predictor; the sum of the points results in a risk score ranging from ≤60 (very low; 30-day mortality <1 percent and one year mortality <10 percent) to >150 (very high; 30-day mortality >50 percent and one year mortality >70 percent).
An on-line calculator for this risk model is available at www.ccort.ca/CHFriskmodel.aspx.
Heart Failure Survival Score — The Heart Failure Survival Score (HFSS) is another prospectively validated model, developed in and for patients with advanced HF (NYHA class III or IV, (table 1) [50]. This score was derived from a multivariable analysis of 268 ambulatory patients referred for consideration of cardiac transplantation and validated in 199 similar patients. The predictors of survival in the HFSS include:
* Presence or absence of coronary artery disease
* Resting heart rate
* LVEF
* Mean arterial blood pressure
* Presence or absence of an interventricular conduction delay on ECG
* Serum sodium
* Peak VO2
In an invasive version of the HFSS, pulmonary capillary wedge pressure is included as an eighth variable. The HFSS stratifies patients into low, medium, and high risk categories, based upon a sum of the variables above multiplied by defined coefficients. Among the patients in the validation sample, one-year survival rates without transplant for these three strata were 88, 60, and 35 percent, respectively.
The HFSS is often used as an aid to selecting patients for cardiac transplantation. However, the above survival rates may not be applicable to current practice since the model was derived before the use of modern therapies that improve survival in patients with HF, including beta blockers, angiotensin II receptor blockers, aldosterone antagonists, implantable cardioverter-defibrillators, and cardiac resynchronization therapy [55]. In addition, some of these therapies directly affect some of the parameters in the risk model (eg, beta blockers and resting heart rate, and cardiac resynchronization therapy and conduction delay). (See "Indications and contraindications for cardiac transplantation", section on 'Heart Failure Survival Score'.)
Seattle Heart Failure Model — The Seattle Heart Failure Model differs from the prior models in two ways [51]:
* The model was derived and validated in a broad HF population, including both general outpatients and advanced HF patients.
* The model incorporates a wide range of readily available clinical variables, including medications and devices.
The model was derived in 1125 advanced HF patients from the PRAISE trial (average LVEF 21 percent, average NYHA class 3.6). None of the patients in the derivation cohorts were treated with beta blockers, aldosterone antagonists, or ICDs, but validation of the model in cohorts from subsequent trials evaluated the impact of these interventions. The model was prospectively validated in five additional cohorts totaling 9942 HF patients and 17,307 person-years of follow-up.
The diversity of the validation cohorts is illustrated by the following:
* Average LVEF ranged from 22 to 35 percent.
* Average NYHA class ranged from 2.2 to 2.9 (table 1).
* Wide ranges of use of beta blockers (24 to 72 percent) and potassium-sparing diuretics (5 to 35 percent).
Across these populations, the model provided an extremely accurate estimate of one-, two-, and three-year survival (r values ranging from 0.97 to 0.99).
The Seattle model also provides information about the likely mode of death. In an analysis of 10,538 ambulatory patients with predominantly systolic HF (NYHA class II to IV), the score was predictive of the risk of sudden death and of pump failure [56]. Compared with patients with a score of 0, the risk of sudden death progressively increased with higher scores, up to a 7-fold higher risk with a score of 4. Conversely, the proportion of deaths caused by sudden death versus pump failure decreased from a ratio of 7:1 with a score of 0 to a ratio of 1:2 with a score of 4.
An online calculator is available at www.SeattleHeartFailureModel.org. The calculator can also estimate the effect of adding new therapies on the risk of mortality.
Effect of medical therapy — Although improved, mortality is still appreciable with appropriate medical therapy. The following results were noted in the treatment arms of the major ACE inhibitor trials (graph 2A-C):
* The mortality rate was 36 percent at one year for patients with NYHA class III or IV disease compared to 52 percent in those not receiving an ACE inhibitor [45]. Similar results were noted in another study of 499 patients with class III or IV HF, 75 percent of whom were receiving an ACE inhibitor and 50 percent of whom were treated with digoxin [57]. The one-year mortality was 35 percent and the rate of death or hospital readmission (2.05 times and 27.6 days per year) was 81 percent.
* The mortality rate was 18 to 20 percent at two years for NYHA class II to III disease compared to 25 percent in those not receiving an ACE inhibitor [46,47].
* Among patients with asymptomatic left ventricular dysfunction, the incidence of cardiovascular death or symptomatic HF was 30 percent at just over three years compared to 39 percent in those not receiving an ACE inhibitor [48].
Further improvements have been made with the addition of beta blockers and spironolactone or eplerenone to ACE inhibition (graph 3A-B). (See "Use of beta blockers in heart failure due to systolic dysfunction" and "Use of diuretics in heart failure".)
The survival benefit associated with ACE inhibitors and hydralazine plus nitrates, although statistically significant, is relatively modest in absolute terms. Most studies have reported a 15 to 20 percent reduction in overall mortality. In the SOLVD trial of New York Heart Association Class II and III HF, for example, enalapril lowered the four-year mortality rate from 42 to 36 percent; this represents an absolute improvement of 6 percent (graph 2B) [46]. Averaging the benefit from these studies through the entire treated population translates into a mean increase in survival of less than six months [34]. Thus, the benefit is apparent only in large populations.
The survival benefit may be somewhat greater with beta blockers. As an example, a meta-analysis of 21 trials involving 5849 patients treated for a median of six months found that beta blockers reduced overall mortality by 39 percent [58]. It was estimated that during the first year this therapy would save 3.8 lives per 100 patients treated [59]. In the MERIT-HF trial of almost 4000 patients with NYHA class II to IV HF who were treated with ACE inhibitors and digoxin, the addition of metoprolol produced a significant reduction in one-year mortality (7.2 versus 11 percent with placebo) (graph 3A) [60]. (See "Use of beta blockers in heart failure due to systolic dysfunction".)
The magnitude of benefit can be illustrated by the marked short-term improvement in survival of patients with NYHA class IV HF. In controlled trials, one-year mortality fell from 52 percent in the CONSENSUS trial published in 1987 in which ACE inhibitors and beta blockers were not used [45] to 11.4 percent in the COPERNICUS trial published in 2001 in which both drugs were used [61]. Although it is not possible to compare these trials directly due to differences in entry criteria, the mortality rate in some groups of patients with advanced HF is now in the range of 15 to 25 percent. The observed improvement is due to the use of HF drugs (ACE inhibitors, beta blockers) and other modalities aimed at underlying diseases (eg, aspirin, statins).
In addition to these specific therapies for HF, better control of hypertension and of coronary risk factors also contribute to the improvement in outcome.
Unfortunately, there are major lags in the appropriate adoption of the drugs that improve survival in HF [34,62,63]. Even among treated patients, the doses used are often lower than those in the clinical trials [63]. (See "Overview of the therapy of heart failure due to systolic dysfunction", section on ACE inhibitors and other vasodilators.)
Effect of demographic factors — The survival of patients with HF is influenced by age, gender, race and the cause of the cardiomyopathy.
Effect of age — The mortality rate in treated patients with HF increases with age [33,64-66]. In a report from the Framingham Study, for example, mortality increased with advancing age (hazard ratio 1.27 and 1.61 per decade in men and women, respectively) [33]. Similar findings were noted in the population-based study from Canada [64]. Patients aged 65 to 74 and ≥75 had an independent increase in one year mortality compared to patients aged 25 to 49 (odds ratio 2.18 and 4.24, respectively).
The prognosis of HF in elderly adults was assessed in a community-based review of over 5500 persons from the Cardiovascular Health Study [67]. Among the 5 percent with HF, the LVEF was normal in 63 percent, borderline low in 15 percent, and overtly reduced in 22 percent. Compared to the mortality rate of 25 deaths per 1000 person-years in those without HF or left ventricular dysfunction, the mortality rates were 87, 115, and 154 per 1000 person-years in those with normal, borderline, and reduced left ventricular function, respectively.
Despite the increase in risk in the elderly, mortality improvements have occurred over time [42,68]. In the report from the Framingham Heart Study mentioned above of subjects between the ages of 65 and 74, five year mortality between 1950-69 and 1990-99 fell from 66 to 54 percent in men and 47 to 40 percent in women who survived at least 30 days after the onset of HF [42].
Effect of gender — The prognosis has generally been better in women than men with HF [33,42,69]. In data from the Framingham Heart Study, the median survival time after diagnosis was 3.2 years in women and 1.7 years in men; after five years, 38 percent of women and 25 percent of men were alive [33]. If only persons who survived the first 90 days were considered, thereby excluding patients with acute myocardial infarction and those requiring early surgery, the five year survival statistics improved to 53 percent in women and 35 percent in men. In the period from 1990 to 1999, five year survival was 60 percent in women compared to 46 percent in men [42].
A reduced risk in women has also been noted in modern therapeutic trials [69-73]. A pooled analysis of five randomized trials testing a variety of therapies among patients with reduced LVEF (PRAISE, PRAISE II, MERIT HF, VEST, and PROMISE) included a total of 8791 men and 2851 women [69]. In multivariable analysis, female gender was associated with significantly longer survival (hazard ratio 0.77).
Similar findings were noted in a comparison of outcomes in 2400 women and 5199 men in the CHARM trial, which included patients with both reduced and preserved LVEF [73]. Women had lower risks of most fatal and nonfatal outcomes; these differences were not explained by LVEF or the cause of heart failure [73].
Effect of race — The effect of race on the prognosis of HF is uncertain since different studies have revealed contrasting findings:
Higher mortality in blacks in a post hoc analysis from the SOLVD trial of enalapril in patients with asymptomatic LV dysfunction or overt HF [74,75]. (See "ACE inhibitors in heart failure due to systolic dysfunction: Therapeutic use", section on Importance of race.)
* Lower mortality in blacks in an analysis of nearly 30,000 Medicare beneficiaries hospitalized for HF in 1998 and 1999 [76].
* No difference in mortality between whites and blacks in a post hoc analysis from the DIG trial of digoxin therapy [77].
Effect of cause of cardiomyopathy — The etiology of HF may be predictive of long-term outcome. (See "Definition and classification of the cardiomyopathies".)
This was addressed by one study of 1230 patients with an initially unexplained cardiomyopathy, which analyzed the outcome based upon the etiology of the cardiomyopathy; after a mean follow-up of 4.4 years, 34 percent of patients died and 4.6 percent underwent cardiac transplantation [78]. Compared to those with an idiopathic cardiomyopathy, which served as the reference group, the following findings were noted (graph 4):
* Survival was better in patients with peripartum cardiomyopathy (hazard ratio 0.31).
* Survival was worse in patients with infiltrative myocardial disease, particularly amyloidosis or hemochromatosis (hazard ratio 7.41 and 8.88, respectively), HIV infection (hazard ratio 5.86), doxorubicin therapy (hazard ratio 3.46), ischemic heart disease (hazard ratio 1.52), or connective tissue disease (hazard ratio 1.75).
* Survival was the same in patients with hypertension, myocarditis, sarcoidosis, substance abuse, or other causes.
In a review of patients from the SOLVD treatment trials, the presence of diabetes had a differential impact on mortality in patients with HF [79]. In adjusted analyses, diabetes significantly increased all-cause mortality in patients with ischemic cardiomyopathy but not those with nonischemic cardiomyopathy (relative risk 1.37 and 0.98, respectively).
Natural history of recent onset IDC — The natural history of recent onset idiopathic dilated cardiomyopathy (IDC) and HF is variable and difficult to predict. Although the one year mortality may be as high as 25 percent, a substantial proportion of these patients improve spontaneously over time. The myocardial contractile response to exogenous catecholamines may be a method to predict those patients who will recover. This hypothesis was tested in a series of 22 patients with recently diagnosed (4±3 months) IDC and a LVEF <40 percent who underwent dobutamine echocardiography; the change in LVEF and left ventricular geometry (sphericity index) in response to dobutamine predicted recovery of left ventricular function [80]:
* Patients with an LVEF >40 percent at six months had an increase of >14 percentage points in baseline LVEF with dobutamine while the improvement was <6 percentage points in those patients with an LVEF that remained <40 percent during follow-up.
* The left ventricular sphericity index in end-diastole on dobutamine, defined as the ratio of left ventricular length (from the apex to the middle of the mitral annular plane) to left ventricular width (at the midpoint of left ventricular length in the four chamber view), correlated with the sphericity index at six months.
Seasonal variation — Several studies have found a seasonal pattern of deaths from myocardial infarction and sudden death, with more fatal events occurring in the winter than the summer. (See "Psychosocial and other social factors in acute myocardial infarction" and "Psychosocial factors in sudden cardiac arrest".) A similar seasonal variation has been seen in men and women with chronic HF [81,82]. In a large study from France, deaths from HF peaked during the winter months of December and January [81]. The distribution of monthly deaths differed by up to 35 percent when January was compared to August, which is when deaths were the lowest. Hospitalizations for HF followed the same seasonal pattern, with a winter-spring predominance (graph 5). Approximately one-fifth of the excess in winter admissions has been attributed to respiratory disease [82].
Circadian rhythm — Sudden death in patients with HF does not follow a circadian rhythm, in contrast to the circadian variation (most deaths between 6 AM and 12 PM) seen in the occurrence of out-of-hospital sudden death or acute myocardial infarction in the general population. An analysis of 1153 patients in the PRAISE trial found a uniform distribution of sudden death in patients with a nonischemic cardiomyopathy, while there was a PM peak in those with an ischemic cardiomyopathy [83]. This PM peak was not altered by the use of antiischemic or antithrombotic therapy. (See "Psychosocial factors in sudden cardiac arrest".)
Diastolic dysfunction — The prognosis of patients with symptomatic HF with normal LV systolic function is less well defined than in those with systolic dysfunction. Data from the Framingham Heart Study, the VHeFT trials, and a community-based survey of elderly subjects from the Cardiovascular Health Study revealed similar findings: diastolic dysfunction was associated with a better prognosis than HF due to systolic dysfunction (annual mortality 8 to 9 versus 15 to 19 percent versus 1 to 4 percent in matched controls) (graph 6) [35,67,84]. Later studies have shown different absolute percentages but a similarly lower risk ratio for diastolic dysfunction. (See "Treatment and prognosis of diastolic heart failure", section on 'Prognosis'.)
CAUSES OF DEATH IN HF — The two main causes of death in patients with HF are sudden or arrhythmic death (defined as death within one hour of the onset of cardiovascular collapse in a previously stable patient) and progressive pump failure (defined as cardiac death with preceding symptomatic or hemodynamic deterioration) [45,46]. In a 38 year follow-up of patients in the Framingham Heart Study, the presence of HF significantly increased overall and sudden cardiac death (SCD) mortality by five-fold (graph 6) [85]. The SCD death potential in men and women with HF was as great as that noted in patients with overt coronary heart disease.
SCD and progressive HF — Published series suggest a relatively consistent pattern with 30 to 50 percent of all cardiac deaths in patients with HF being categorized as sudden deaths, with or without preceding symptoms [38,46,48,60,85,86]. However, it is often difficult to distinguish those dying suddenly and unexpectedly from those experiencing terminal arrhythmias in the setting of progressive hemodynamic deterioration [87-89].
As a result, clinical studies have classified deaths in differing ways. In the AIRE trial, 45 percent of patients who died suddenly had severe or worsening HF prior to their death; only 39 percent of sudden deaths were thought to be due to arrhythmia [88]. In the control arm of the MADIT II trial, of 72 deaths considered to be cardiac, 31 percent occurred within one hour of symptom onset, 36 percent occurred greater than one hour after symptom onset, and 33 percent were unwitnessed [89]. In the ICD arm, defibrillator therapy reduced the incidence of deaths occurring within one hour of symptoms and unwitnessed deaths, suggesting that these groups included most of the patients with terminal arrhythmia.
It has been suggested that progressive pump failure, sudden death, and sudden death during episodes of clinical worsening each account for approximately one-third of deaths [87]. Ventricular tachycardia degenerating into ventricular fibrillation is the most common cause of sudden death; a bradyarrhythmia or pulseless electrical activity (electromechanical dissociation) is responsible in 5 to 33 percent of cases [87], although the frequency may be increased in patients with advanced HF [90].
The three major classes of drugs that improve survival in HF - angiotensin converting enzyme (ACE) inhibitors, beta blockers, and aldosterone antagonists - have different effects on the causes of death. The benefit of ACE inhibitors is primarily derived from prevention of progressive myocardial dysfunction, rather than prevention of SCD [45,46,48,86]. Although ramipril significantly reduced sudden death by 30 percent in the AIRE trial, most of these deaths as mentioned above were due to severe or worsening HF rather than arrhythmia [88]. In contrast, beta blockers and, in patients with class IV HF, spironolactone reduce both SCD and death from progressive HF [58,60,91]. (See "ACE inhibitors in heart failure due to systolic dysfunction: Therapeutic use" and "Use of beta blockers in heart failure due to systolic dysfunction" and "Use of diuretics in heart failure".)
The prognostic significance of ICD shocks in patients with HF is discussed separately. (See "Secondary and primary prevention of sudden cardiac death in heart failure and cardiomyopathy", section on 'Prognostic significance of ICD shocks'.)
Role of ischemia — Although sudden death is most often the result of a ventricular tachyarrhythmia, the role of an acute coronary event in these patients may be underestimated. The prevalence of an acute coronary finding (coronary thrombus, ruptured plaque, or myocardial infarction) and its relationship to sudden death was examined in an autopsy study of 171 patients with HF [92]. In patients with significant coronary artery disease, an acute coronary finding was found in 54 percent who died suddenly and in 32 percent who died of myocardial failure, although an acute coronary event had not been clinically diagnosed before death. In contrast, an acute coronary finding was uncommon in those without coronary disease, present in only 5 percent of those dying from sudden death and 10 percent of those dying from myocardial failure.
Reference:
http://www.uptodate.com/home/index.html
|
