Tsuyuki RT, McKelvie RS, Arnold JMO, Avezum, Jr Á, Barretto ACP, Carvalho ACC, Isaac DL, Kitching AD, Piegas LS, Teo KK, Yusuf S. Acute Precipitants of Congestive Heart Failure Exacerbations. Arch Intern Med. 2001;161(19):2337–2342. doi:10.1001/archinte.161.19.2337
Few studies have prospectively and systematically explored the factors that acutely precipitate exacerbation of congestive heart failure (CHF) in patients with left ventricular dysfunction. Knowledge of such factors is important in designing measures to prevent deterioration of clinical status. The objective of this study was to prospectively describe the precipitants associated with exacerbation of CHF status in patients enrolled in the Randomized Evaluation of Strategies for Left Ventricular Dysfunction Pilot Study.
We conducted a 2-stage, multicenter, randomized trial in 768 patients with CHF who had an ejection fraction of less than 40%. Patients were randomly assigned to receive enalapril maleate, candesartan cilexetil, or both for 17 weeks, followed by randomization to receive metoprolol succinate or placebo for 26 weeks. Investigators systematically documented information on clinical presentation, management, and factors associated with the exacerbation for any episode of acute CHF during follow-up.
A total of 323 episodes of worsening of CHF occurred in 180 patients during 43 weeks of follow-up; 143 patients required hospitalization, and 5 died. Factors implicated in worsening of CHF status included noncompliance with salt restriction (22%); other noncardiac causes (20%), notably pulmonary infectious processes; study medications (15%); use of antiarrhythmic agents in the past 48 hours (15%); arrhythmias (13%); calcium channel blockers (13%); and inappropriate reductions in CHF therapy (10%).
A variety of factors, many of which are avoidable, are associated with exacerbation of CHF. Attention to these factors and patient education are important in the prevention of CHF deterioration.
DESPITE MANY advances in congestive heart failure (CHF) therapy, worsening of clinical status, often to the point of requiring hospitalization, remains a common occurrence.1 It is estimated that there are 700 000 hospitalizations for CHF per year in the United States. In the Studies of Left Ventricular Dysfunction registry of 6273 consecutive patients with CHF,2 19% were hospitalized 1 or more times for CHF and 27% died or were hospitalized for CHF within 1 year. In 2 large CHF trials, the Studies of Left Ventricular Dysfunction Treatment Trial3 and the Digitalis Investigation Group Trial,4 26% and 27%, respectively, of patients in the active treatment groups were hospitalized at least once during the 3 to 4 years of follow-up. Indeed, approximately two thirds of the economic burden of CHF is due to hospitalizations for worsening clinical status.5
Clinical and lifestyle factors associated with deterioration of patients with CHF have not been systematically addressed. Most studies that have explored the prognostic importance of CHF have relied on historical, clinical, or investigational data obtained at "entry" into the study. The impact of these prognostic indicators on later deterioration of clinical status, whether transient or permanent, were then assessed at a remote time point.6- 8 However, given the magnitude of the problem, knowledge of the immediate precipitants of CHF exacerbation, particularly if avoidable, provides useful insight for clinicians in preventing the deterioration. Therefore, the purpose of this study was to prospectively and systematically explore the immediate precipitants associated with exacerbation of CHF in patients enrolled in a 43-week multicenter clinical trial.
The precipitants of CHF exacerbations were determined within the context of the Randomized Evaluation of Strategies for Left Ventricular Dysfunction Pilot Study. The purpose, methods, and main results of this double-blind, randomized, 2-stage, multicenter trial in 768 patients with symptomatic CHF (New York Heart Association functional class [NYHA-FC] II-IV), an ejection fraction of less than 40%, and a 6-minute walk distance of less than 500 m have been described previously.9- 11 In stage I, patients were randomized to receive the angiotensin II receptor blocker candesartan cilexetil alone (4, 8, or 16 mg daily), candesartan cilexetil plus enalapril maleate (4 or 8 mg of candesartan cilexetil daily plus 10 mg of enalapril twice daily), or enalapril alone (10 mg twice daily). Patients were followed up for 17 weeks, at which time they were considered for a second randomization to receive controlled-release metoprolol succinate (200 mg daily) or placebo (stage II). Doses of all medications were titrated upwards to the target doses in 6 to 8 weeks. Patients were followed up in stage II for 26 weeks, for a total study duration of 43 weeks. Patients not randomized in stage II (usually because of β-adrenergic blocking agent contraindications or patient refusal) continued to receive candesartan, enalapril, or both and were also followed up for another 26 weeks. Study end points were 6-minute walk distance (primary end point), neurohormone levels, ejection fraction, cardiac volumes, symptoms, quality of life, and safety. All end points were measured at baseline and at the end of stage I and stage II. Written informed consent was obtained from all study participants, and all study centers received local research ethics committee approval.
As part of the clinical event reporting for the study, investigators completed a CHF Event Form for all patients experiencing any worsening of CHF. This form was completed regardless of whether the patient required hospitalization. Data collected on the form included symptoms and physical signs present at the time of diagnosis of the CHF exacerbation, treatments used, and factors considered by the investigator to have contributed to the episode. Investigators also indicated the single most responsible factor for the patient's deterioration. For the purposes of the present analysis, each episode of worsening CHF was treated as an independent event.
The demographic and clinical characteristics of patients in the Randomized Evaluation of Strategies for Left Ventricular Dysfunction Pilot Study who had 1 or more CHF events (episodes of worsening of CHF) compared with those who did not are shown in Table 1. A total of 180 patients experienced a total of 323 CHF events during the 43-week trial. Patients who experienced CHF events were of similar sex distribution (approximately 83% were men) and age (approximately 63 years) as those who did not. Patients in both groups also had a similar duration (approximately 80% had a duration of >12 months) and cause (approximately 70% ischemic) of CHF. More patients with a CHF event were in NYHA-FC III or IV at baseline (51%) compared with those without an event (29%).
There were no striking differences in clinical examination findings at baseline in patients with vs without a CHF event, with the exception of more peripheral edema and jugular venous distention in patients with a subsequent CHF event (Table 2).
Nonstudy medications taken by patients in the 48 hours before the CHF exacerbation are shown in Table 3. Loop diuretics were received by 92% of patients, digoxin by 76%, acetylsalicylic acid by 52%, nitrates by 47%, and open-label β-adrenergic blocking agents by 14%.
Of the 323 CHF events, 143 events required that the patient be hospitalized, and 180 cases were managed on an ambulatory basis (Table 4). Five patients died in the hospital. Patients gained an average weight of 1.16 kg compared with baseline. Physical examination revealed lung crackles in 66% of cases, peripheral edema in 52%, elevation of jugular venous pressure in 55%, and a third heart sound in 36%. Hospitalized cases had more crackles (76% vs 56%) and were more likely to have a chest radiograph performed compared with those not hospitalized (76% vs 21%). Treatments administered for the exacerbation of CHF included oral diuretics in 56% of cases, intravenous diuretics in 48%, and addition of a new diuretic in 19%. Intravenous inotropes were administered in 16% of cases and intravenous digoxin in 10%. Nitrates were started or increased in 9% of cases. Mechanical ventilation and intra-aortic balloon pump were used rarely. As expected, hospitalized patients received intravenous therapy such as diuretics, inotropes, digoxin, and vasodilators more frequently, although 22% of ambulatory patients also received intravenous diuretics.
Factors that the investigators believed to be associated with exacerbation of CHF are shown in the "Total" column of Table 5 (items are not mutually exclusive). The most commonly identified causes of CHF exacerbation were excessive salt intake (22%), other noncardiac disorders in 20%, study medications (chiefly metoprolol, as indicated by the investigators) in 15%, use of antiarrhythmic agents in the past 48 hours in 15%, development of arrhythmias in 13%, calcium channel blockers in 13%, and inappropriate reductions in CHF therapy in 10%. Other precipitants were identified in 46% of cases and included a variety of potential causes. A review of the investigators' free-text comments in the latter and "other noncardiac" categories indicated that upper respiratory tract infections, especially pneumonia, accounted for 37 events (11% overall). Noncompliance with medications, uncontrolled hypertension, and coronary ischemia were implicated in only 7%, 2%, and 2%, respectively. There were no major differences observed in the factors associated with CHF exacerbation between patients who were hospitalized vs not hospitalized.
The most common primary causes of CHF exacerbations identified by the investigators were excessive salt intake in 15%, other noncardiac disorders in 15%, and inappropriate reductions in CHF therapy in 9% (Table 5). The primary precipitant was indicated as "other noncardiac" or "other" in 15% and 38% of cases, respectively. A review of the free-text summaries documented in these categories indicated respiratory tract infection in 31 cases (10% overall), study metoprolol in 16 (5% overall), and excessive fluid intake in 12 (4% overall). Noncompliance with medications, coronary ischemia, uncontrolled hypertension, arrhythmias, and other medications were not commonly indicated as primary causes of worsening CHF.
Congestive heart failure is one of the most common reasons for hospital admission.12 Yet, for such a clinically and economically important condition, there are only limited data describing the precipitants of CHF exacerbation. Careful attention to these factors might have an important impact in preventing clinical deterioration in these patients.
This study prospectively collected information on acute precipitants, physical findings, and medications used immediately before and during treatment for the episodes of acute exacerbation and showed that a variety of potentially preventable factors led to the decompensation of patients with CHF. Except for an excess of patients with NYHA-FC III and IV presenting with subsequent exacerbations (as one might expect), baseline clinical characteristics and physical findings did not show obvious differences between patients who subsequently developed CHF events and those who did not. Thus, baseline variables are not helpful in predicting subsequent events in this group of patients with characteristics similar to that of the Studies of Left Ventricular Dysfunction Treatment Trial in terms of sex distribution, age, proportion of patients with an ischemic origin of CHF, and NYHA-FC.13
The medications taken by patients in the 48 hours before the exacerbation were as expected for this type of patient, with diuretics, digoxin, acetylsalicylic acid, and nitrates being used commonly. One major difference between the present study population and unselected patients with CHF is that all study patients were receiving an angiotensin-converting enzyme inhibitor or an angiotensin II antagonist compared with only about half in unselected patients with CHF.14,15 In addition, most patients received the study β-adrenergic blocking agent, its placebo, or an open-label β-adrenergic blocking agent.
Patients presenting with worsening of CHF symptoms had the typical physical signs of lung crackles, edema, and elevation of jugular venous pressure. At least some weight gain was associated with most CHF exacerbations. The weight gain observed was small (1.16 kg), which might be indicative of the fact that many patients do not weigh themselves daily (and those who do, experience less weight gain). This simple, important measurement should be used by patients to help identify the onset of an impending exacerbation of their disease, and they should seek medical attention before severe cardiovascular compromise occurs.
As expected, in-hospital treatments administered for CHF exacerbation required adjustment of therapies, including oral and intravenous diuretics; initiation or adjustment of vasodilators; initiation of inotropic therapy; and, for a few, mechanical ventilation and intra-aortic balloon pump. The nonhospitalized group also had adjustments in diuretic and vasodilator therapies. Twenty-two percent of nonhospitalized patients received intravenous diuretics, which might be an important strategy to keep patients out of the hospital.
The most common factors contributing to CHF exacerbation include excessive salt intake due to lack of knowledge of, or failure to comply with, salt restriction; other miscellaneous noncardiac disorders; use of inappropriate medications (antiarrhythmic agents, calcium channel blockers, or inappropriate reductions in other CHF medications); and development of arrhythmias (primarily tachyarrhythmias). It is notable that many of these precipitants are avoidable. Study medications, particularly the β-adrenergic blocking agent, or placebo were cited as the causes of acute exacerbation in 15% of cases. Although long-term β-blockade has now been conclusively demonstrated to improve mortality and morbidity rates in patients with CHF, it is not clear whether β-adrenergic blocking agents would account for this proportion of acute exacerbations if the patients were not in a masked study (ie, half of the patients randomized in stage II received a placebo β-adrenergic blocking agent). Taken together, these factors suggest that measures to prevent worsening of CHF should include patient education regarding salt intake (preferably by a dietitian) and cautious use or modification of medications that have negative inotropic properties. In reviewing the free-text descriptions provided by the investigators, pulmonary infectious processes were commonly implicated in CHF exacerbation. At least some of these events would be potentially preventable with the use of pneumococcal or influenza vaccine. A recent study has shown, however, that pneumococcal vaccine is underused in this high-risk population.16
Our finding that medication noncompliance, myocardial ischemia, uncontrolled hypertension, arrhythmia, and the effects of other medications were not common causes of worsening CHF in this patient population differs somewhat from findings from other studies. Chin and Goldman,6 in 1997, reported on causal factors in 435 patients admitted to the hospital for CHF. In this study, a single investigator retrospectively reviewed the medical records of patients admitted to the hospital for CHF, looking for possible precipitants of the event. The most commonly associated reasons for clinical deterioration were acute anginal chest pain (33%), respiratory tract infection (16%), uncontrolled hypertension (15%), and noncompliance with medications (15%) (these categories are not mutually exclusive). The major limitation of this study was the retrospective nature of the data collection, which depends on the completeness of medical charting. Information on lifestyle factors might not be systematically documented. As well, this type of retrospective analysis cannot accurately determine the temporality or causality of the precipitants, ie, whether the precipitant was a cause or an effect of CHF exacerbation (eg, ischemia might cause worsening of CHF symptoms but might also be a result of it). Ghali et al7 reported on precipitating factors leading to decompensation of CHF in 101 patients admitted to Cook County Hospital, Chicago, Ill. Patients were prospectively evaluated through a systematic patient interview and medical chart review. The most common precipitating factors identified were lack of compliance with diet, drugs, or both in 64.4%; uncontrolled hypertension in 43.6%; and cardiac arrhythmias in 28.7% (these categories are not mutually exclusive). The authors pointed out that their study population included mostly individuals of low socioeconomic status and that their results might not be applicable to other patient populations. Opasich et al8 described patients who were referred to the CHF service at an Italian hospital in 2 years. In their prospective analysis of potential causative factors in 328 nonfatal decompensations in 161 patients, the presence of arrhythmias in 24%, infections in 23%, poor compliance in 15%, and angina in 14% were identified (these categories are not mutually exclusive). The investigators did not review cases of decompensation in patients who died because of a stated lack of information surrounding the event, which might somewhat limit the applicability of the results. It is also possible that patients referred to their CHF service are not representative of the general CHF population.
The present study may have underestimated the importance of poor compliance with CHF medications for 2 reasons. First, the patients participating in this study were enrolled in a clinical trial. Patients who volunteer to participate in clinical trials may be more interested in their disease state, and therefore may be more compliant with their medications. Second, it is known that patient interview or clinician impression is notoriously insensitive to detect poor compliance.17 Therefore, poor compliance may well be an important and often unrecognized factor in CHF exacerbation, and clinicians should consider the use of alternative data sources, such as pharmacy records, in patients with poor control of symptoms.
The main strengths of the present study are that it was performed prospectively, used a systematic approach to elucidate the precipitants of CHF, and was conducted in 60 clinical centers using a well-defined cohort of patients. Documentation of the precipitants of CHF in the present study was performed prospectively using a standardized approach by the study physicians and research nurses who had been following these patients closely rather than by a third party. Presumably, study personnel might have had more insight into the patient's condition antecedent to the exacerbation and thus provided a more accurate assessment of causality. Another unique feature of the present study is that it reports on all CHF exacerbations, including those requiring and not requiring hospitalization. This might be important because knowledge of the acute precipitants of CHF in all patients (not just those hospitalized) might help prevent hospitalization.
As part of the study protocol, all patients received an angiotensin II antagonist, an angiotensin-converting enzyme inhibitor, or both. Indeed, more than 90% of patients reached their target dose. This might differ somewhat from the real-world situation because many patients do not receive either of these therapies14,15 or receive suboptimal doses.18 The Studies of Left Ventricular Dysfunction Treatment Trial3 showed a 26% relative risk reduction in death or hospitalization for CHF with enalapril therapy compared with use of placebo. Therefore, lack of angiotensin-converting enzyme inhibitor use might be a significant (and preventable) cause of CHF exacerbation; however, because of the design of the present study, the contribution of this important variable could not be assessed.
Heart failure exacerbations were identified by the study investigators. Although the study was regulated by trained research monitors, it is possible, but unlikely, that some patients with worsening of symptoms (particularly those not requiring hospitalization) were missed. The precipitants of CHF exacerbation in the present study were identified within the context of a clinical trial, which raises the issue of patient selection bias, particularly for those most likely to be compliant with medical advice.
Deterioration of clinical status is a common occurrence in patients with CHF. Such patients are clinically fragile, and a variety of preventable factors can lead to deterioration, such as excessive salt intake and use of negative inotropic medications. Ensuring that all patients with CHF receive vaccination for influenza and pneumococcus might also reduce these causes of CHF worsening. Health care providers involved in the care of these patients should be aware of these factors and should take steps to incorporate this information into the care and education of their patients.
Accepted for publication March 13, 2001.
Corresponding author and reprints: Ross T. Tsuyuki, BSc(Pharm), PharmD, MSc, Epidemiology Coordinating and Research (EPICORE) Centre, Division of Cardiology, University of Alberta, 213 Heritage Medical Research Centre, Edmonton, Alberta, Canada T6G 2S2 (e-mail: firstname.lastname@example.org).