- Email: [email protected]
Appropriateness of digoxin use in medical outpatients
Appropriateness of Digoxin Use in Medical Outpatients MAXINEA. PAPADAKIS, M.D., BARRYM. MASSIE,M.D.san~ranck~o,
Ca/ifornia
PURPOSE:Digoxin is the third most commonly prescribed drug, yet limited information exists about its use in outpatients. Therefore, 242 medical outpatients receiving digoxin at our hospital were studied to evaluate the appropriateness of its use, defined by: (1) current or past supraventricular arrhythmias and/or (2) left ventricular systolic dysfunction (ejection fraction less than 45 percent). PATIENTSANDMETHODS: Charts of242 patients receiving digoxin were obtained. The patients were divided into groups based upon their physician’s stated indication for digoxin therapy. Patients with only a clinical diagnosis of congestive heart failure (CHF) underwent echocardiography or radionuelide angiography to quantify left ventricular systolic function. Those with documented supraventricular arrhythmias and/or those with confirmed left ventricular systolic dysfunction were classified as appropriate candidates for digoxin. RESULTS:Ninety-five percent of patients received digoxin for appropriate indications; 75 percent had confirmed supraventricular arrhythmias (27 percent also had CHF) and 20 percent with normal sinus rhythm had documented systolic dysfunction. However, physicians had difficulty in the clinical assessment of left ventricular function; 18 percent of patients with sinus rhythm and CHF by the Framingham scoring system and 20 percent of those with supraventricular arrhythmias and CHF had preserved systolic function, An S3 was present in 15 percent of patients with preserved ejection fraction and CHF and in 69 percent with low ejection fraction; hypertension was significantly more common in the former group. Noninvasive assessment of systolic function was obtained in 97 percent of patients independent of this study, yet some patients without supraventricular arrhythmias and with documented preservation of systolic function continued to receive the drug. CONCLUSION: Noninvasive assessment of left ventricular function, which appears to have become routine, is of value in the appropriate utilization of digoxin, since clinicians’ assessment of left ventricular function may be inaccurate. Physicians also do not always discontinue digoxin therapy when indicated.
From the Divisions of General Internal Medicine and Cardiology, San Francisco Veterans Administration Medical Center and the University of California, San Francisco, San Francisco, California. Requests for reprints should tration Medical Center, 4150Clement fornia 94121. Manuscript submitted
Street, March
111 Al, San Francisco, Cali18, 1988. and accepted in
D
igoxin is the third most commonly prescribed drug in the United States [l]. There are two major indications for its use: control of the ventricular response in supraventricular tachyarrhythmias and improvement of myocardial performance in congestive heart failure (CHF) secondary to dilated cardiomyopathy. A number of studies have questioned the efficacy of digoxin for patients with CHF and sinus rhythm, but several of these have included subjects without documented left ventricular systolic dysfunction [26]. In recent years, the frequent occurrence of symptoms and signs of CHF due primarily to diastolic dysfunction and episodic CHF due to ischemic left ventricular dysfunction has been recognized [7]. Response to digoxin would not be expected in such patients. To our knowledge, only one previous study has examined the prevalence of inappropriate usage of digoxin in outpatients [8]; 42 percent of patients were found to be receiving the drug for a questionable reason. Our study was designed to answer the following questions: (1) Do internists prescribe digoxin appropriately based on clinical criteria? (2) Of those patients appropriately prescribed digoxin clinically, how many are potentially benefiting from the drug based on demonstration of resting systolic dysfunction? (3) Is noninvasive assessment of left ventricular function indicated for outpatients receiving digoxin for clinical findings of CHF?
PATIENTS AND METHODS Setting The San Francisco Veterans Administration Medical Center (SFVAMC) is a primary teaching hospital of the University of California, San Francisco School of Medicine. All faculty and house staff hold university appointments. The SFVAMC Group Practices (general medical clinics) are staffed by 24 categorical Internal Medicine house staff, six Division of General Internal Medicine faculty, and three nurse practitioners. Patients Three thousand patients are enrolled in the SFVAMC Group Practices. Since the Veterans Administration health care system provides patient medication, it was possible to identify those receiving digoxin. Pharmacy records from September 1, 1986, of all Group Practice patients were reviewed. Charts were available for 242 of 245 patients receiving digoxin; they constitute the sample for this study. Protocol The protocol for this study was approved by the University of California, San Francisco, Human and Environmental Protection Committee. The following data were abstracted from hospital
September
1988
The American
Journal
of Medicine
Volume
85
365
DIGOXIN
USE IN OUTPATIENTS
/ PAPADAKIS
AND MASSIE
ferred for echocardiography or, if the result was technically unsatisfactory, radionuclide angiography, to quantify left ventricular function. The patients were divided into the following three groups based upon their physician’s stated indication for digoxin therapy. GROUP 1. Patients with current or past supraventriculaf arrhythmias receiving digoxin for control of ventricular response. GROUP 2. Patients with normal sinus rhythm receiving digoxin for the treatment of CHF. GROUP 3. Patients with supraventricular arrhythmias and the clinical diagnosis of CHF.
TABLEI Demographic Data and Indication for Digoxin Use in 242 Study Patients Number Male sex Mean age (years) Mean length of digoxin use (years) Group 1 (supraventricular arrhythmias) Group 2 (CHF) Group 3 (CHF and arrhythmias) All values expressed
as f standard
deviation
(percent)
240 (99) 70f 11* 4f5 !Y $ij 65 (27) unless stated
Study Criteria for Appropriate Digoxin Use Patients with documented supraventricular arrhythmias and those with confirmed left ventricular systolic dysfunction (ejection fraction less than 45 percent by echocardiography, radionuclide angiography, or cardiac catheterization) were classified as appropriate candidates for digoxin. Although in part arbitrary, this value represents the lower 95 percent confidence limits for a normal echocardiographic ejection fraction in our laboratory and a liberal value for the degree of left ventricular systolic dysfunction usually associated with clinical CHF [lo-121.
TABLEII Appropriateness of Digoxin Therapy Based on Physician’s Indication for Use Number Prescribed (percent) Group 1 (suoraventricular krhythmias) Group 2 (CHF) Group 3 (CHF and arrhythmias)
Percent with Preserved Ejection Fraction
Percent Inappropriate 0
116 (48) 61(25) 65 (27)
18 0
:i
TABLEIll Comparison of CHF Patients with Decreased and Preserved Ejection Fraction (EF)
SB
Hypertension
Preserved EF
30%
15%’
58%1
Dir$nI,‘,“,EF (n = 102)
90%
69%
30%
I p
and clinic chart review: patient’s age, sex, cardiac history, initiation date, and indication for digoxin therapy. Results of physical examination, electrocardiography, and chest radiography at the time of institution of digoxin therapy were noted as well as findings on the most recent physical examinations. Data from objective assessment of left ventricular function by echocardiography, radionuclide techniques, hemodynamic measurements, and left ventricular angiography were obtained from the charts and respective laboratories. Tests whose results were equivocal were reviewed for a definitive interpretation by an observer unaware of the clinical presentation. Clinical CHF Scoring System For patients who had the diagnosis of CHF, its severity was quantified by the investigators based on the scoring system developed by the Framingham study group [9]. The presence of an SBgallop was also noted. Prospective Studies Patients with only a clinical diagnosis of CHF (with or without supraventricular arrhythmias) were re-
366
September
1988
The American
Journal
of Medicine
Volume
Analysis Proportions were compared with a chi-square test and statistical significance was set at the 0.05 level.
RESULTS Demographic data and indications for digoxin use in the 242 study patients are shown in Table I. Most were older men who had received digoxin for several years, reflecting the population served by the SFVAMC. Seventy-five percent of patients were receiving digoxin for control of supraventricular arrhythmias, 27 percent with and 48 percent without CHF. CHF in normal sinus rhythm was the indication for the drug in the remaining 25 percent of patients. The patients’ health care providers had independently performed noninvasive assessment of left ventricular function in nearly all patients with the clinical diagnosis of CHF. Ninety-seven percent of Groups 2 and 3 patients had a mean of 2.2 tests (range, one to 12). Eighteen percent of Group 2 and 20 percent of Group 3 patients had an ejection fraction greater than 45 percent by noninvasive assessment (Table II). Of those with a preserved ejection fraction, less than one third had CHF by the Framingham clinical scoring system (Table III). The mean ejection fraction was 29 percent in the remainder of Groups 2 and 3 patients, and 90 percent of them had CHF by the Framingham scoring criteria. Of patients with a diagnosis of CHF and a normal ejection fraction on their most recent evaluation, only one had a previous abnormal ejection fraction. An S3 was reported as present in 15 percent of patients with preserved ejection fraction and CHF, compared with the majority (69 percent) of those with low ejection fraction; the sensitivity and specificity of SS for decreased left ventricular ejection fraction was only 69 percent and 85 percent, respectively. The most common accompanying diagnosis in patients with the diagnosis of CHF and preserved ejection fraction was hypertension, which was present in
85
DIGOXIN
58 percent (Table III). Significantly fewer patients (30 percent) with a reduced ejection fraction had a history of hypertension.
COMMENTS Ninety-five percent of medical outpatients at the SFVAMC were receiving digoxin for appropriate indications based upon predetermined criteria derived from the literature. This included 75 percent who had confirmed supraventricular arrhythmias and 20 percent with normal sinus rhythm but with documented left ventricular systolic dysfunction. This percentage compares favorably with that in previous studies by Carlson et al [8] and Lee et al [13], which demonstrated that more than 40 percent of patients were receiving the drug for questionable indications. The patient populations in those studies differed from ours, however. More of our patients had supraventricular arrhythmias as an indication for digoxin therapy than in the study of Carlson et al, and only those with sinus rhythm were evaluated by Lee and colleagues. Our greater number of patients with supraventricular arrhythmias may serve to increase the proportion of patients receiving the drug appropriately. The major problem with inappropriate digoxin use in our study was difficulty in the clinical assessment of left ventricular function; 18 percent of patients receiving digoxin for CHF with sinus rhythm and 20 percent of patients with supraventricular arrhythmias and the diagnosis of CHF had preserved systolic function. This finding is important because patients with sinus rhythm together with preserved systolic function probably do not benefit from digoxin, so this treatment carries unnecessary risk. Those with preserved left ventricular systolic function and supraventricular arrhythmias but with the diagnosis of CHF may have been candidates for alternative or concomitant treatment with verapamil or beta blockers for optimal rate control and other indications, such as hypertension. The efficacy of digitalis as a positive inotropic agent in patients with CHF continues to be debated [14,15]. The issue assumes particular importance because of the potential toxicity of digoxin and the recent availability of alternative effective agents for CHF. In one of the few controlled trials of digoxin use, a beneficial effect of the drug in outpatients was shown, especially in those with a ventricular gallop (Sa) or systolic dysfunction. However, 44 percent of the patients had no appreciable clinical improvement with the drug [13]. Importantly, this study suggests that the subset of patients who clinically benefit from digoxin have chronic systolic dysfunction, rather than episodic symptoms of CHF or diastolic dysfunction. Furthermore, patients with normal left ventricular ejection fraction did not benefit despite their symptoms of CHF. Indeed, many of the studies that have questioned the efficacy of digoxin have included patients with relatively preserved left ventricular systolic function or have not included such measurements. A recently published report comparing digoxin, captopril, and placebo in mild to moderate CHF [16] confirmed the value of digoxin in at least a subset of patients with mild to moderate symptoms and ejection fractions below 40 percent, but did not show an advantage over captopril. A large study comparing digoxin and milrinone in patients with moderate symptoms recently re-
USE IN
OUTPATIENTS / PAPADAKIS AND MASSIE
ported in preliminary form indicated that digoxin was particularly effective in the subset of patients with ejection fractions less than 20 percent [17]. These reports emphasize the importance of appropriately choosing candidates for digoxin, and the requirement for underlying moderate or severe systolic dysfunction for clinical efficacy. Recent studies employing noninvasive techniques have shown that CHF is often associated with relatively preserved systolic function, and that these patients may be clinically indistinguishable from those with severe contractile dysfunction [11,12]. In studies of patients with classic signs and symptoms of CHF referred for radionuclide angiography, approximately 40 percent did not have significantly abnormal systolic function [11,12]. We found that 18 percent of patients receiving digoxin with a diagnosis of CHF had preserved left ventricular ejection fraction. Explanations for this observation include the following: (1) The diagnosis of CHF is incorrect. (2) CHF is caused by predominantly diastolic, rather than systolic, dysfunction. (3) CHF has resolved. (4) Systolic function has normalized with digoxin therapy. If any one of the aforementioned first three reasons is correct, digoxin should be discontinued. Yet despite documentation of preserved ejection fraction in our study, health providers failed to discontinue the drug. Potentially, patients could receive digoxin indefinitely after treatment is instituted. Even if initiation of the drug were begun by another physician, the current physician should reassess whether a continued need exists; if the initial or current indications for the drug are unclear, it should be discontinued. Although digoxin can produce a modest increase in ejection fraction (less than 5 percent), it is rarely of the magnitude to restore the fraction to the normal range [16]. Hence, normal left ventricular function by noninvasive testing in patients with normal sinus rhythm is adequate justification to discontinue digoxin in most patients. One potential exception would be patients with significant mitral regurgitation whose ejection fraction may be inappropriately high in relation to the degree of myocardial dysfunction. The presence of a left ventricular Sa has been considered one of the best predictors of therapeutic response to digoxin, being both very sensitive and specific [13]. Surprisingly, when contrasted to therapeutic response to digoxin, an Ss is much less sensitive but still very specific for decreased left ventricular ejection fraction [13]. Whereas an Sa can occur normally in younger persons with rapid left ventricular filling and in mitral insufficiency, its pathogenesis in left ventricular failure is probably related to the sudden limitation in the filling movement of the left ventricle during diastole [18-201. Yet our study shows that, in clinical practice, detection of an S3 gallop is neither sensitive nor specific for decreased ejection fraction. We interpret these findings as demonstrating that cardiac auscultation by the clinician is not as accurate as that by the investigators, whose observations were verified by phonocardiography or apexcardiography; cardiac auscultation remains a subjective and variable method. Others have noted the presence of an SS in patients with intact systolic function but with the diagnosis of CHF. An Sawas heard in more than 40 percent of older patients with normal ejection fraction and absence of
September
1988
The American
Journal
of Medicine
Volume
85
367
DIGOXIN
USE IN OUTPATIENTS
/ PAPADAKIS
AND MASSIE
hemodynamically significant valvular heart disease, but who met the Framingham criteria for CHF [ll]. Clinical Implications
Noninvasive assessment of left ventricular function appears to have become routine; nearly all patients in this study had received such a test, but physicians sometimes failed to act on the information. On the basis of our demonstration of the inaccuracy of the physical examination for reduced systolic function, we recommend noninvasive assessment of left ventricular function be undertaken in patients diagnosed with congestive heart failure, especially if sinus rhythm is normal. The cost of two-dimensional echocardiography or radionuclide angiography is $331* and $460*, respectively. Although a year’s supply of digoxin is relatively inexpensive ($86*), additional costs include monitoring of therapeutic drug levels ($67.25 per ievel*), physician office visits, and potential drug toxicity. The therapeutic-to-toxic ratio of digitalis is quite narrow, and its toxicity is one of the most prevalent adverse drug reactions encountered in clinical practice [21]. Toxicity may develop in as many as 5 to 15 percent of hospitalized patients receiving digitalis [21]. Equally important, if no impairment of ejection fraction can be demonstrated, digoxin should be withheld or discontinued. * University
of California,
San Francisco
outpatient
charges,
1988.
REFERENCES l.Top 200 drugs of 1987. American Druggist. February 1988; 197: 36-52. 2. Fleg JL, Gottlieb SH, Lakatta EG: Is digoxin really important in treatment of compensated heart failure? Am J Med 1982: 73: 244-250. 3. Gheorghiade M, Belier G: Effects of discontinuing maintenance digoxin therapy in patients with ischemic heart disease and congestive heart failure in sinus rhythm. Am J Cardiol 1983; 51: 1243-1249. 4. Hull SM, Mackintosh A: Discontinuation of maintenance digoxin therapy in gen-
368
September
1988
The American
Journal
of Medicine
Volume
eral practice. Lancet 1977; II: 1054-1055. 5. Johnston GD, McDevitt DG: Is maintenance digoxin necessary in patients with sinus rhythm. Lancet 1979; II: 567-570. 6. Smith Tw: Medical treatment of advanced congestive heart failure; digitalis and diuretics. In: Braunwald E, Moth MB, Watson JT, eds. Congestive heart failure. New York: Grune & Stratton, 1982; 261-278. 7. Parker JO, Ledwich JR, West RO, Case RB: Reversible cardiac failure during angina pectoris: hemodynamic effects of atrial pacihg in coronary artery disease. Circulation 1969; 39: 745-757. 8. Carlson KJ, Lee DC, Goroll AH, Leahy M, Johnson RA: An analysis of physicians’ reasons for prescribing long-term digitalis therapy in outpatients. J Chronic Dis
1985; 38: 733-739. 9. McKee PA, Castelli WP, McNamara PM, Kannel WB: The natural history of congestive heart failure: the Framingham study. N Engl J Med 1971; 285: 1441-
1446. 10. Massie BM. Kramer BL, Gertz EW, Henderson SG: Radionuclide measurement of left ventricular volume: comparison of geometric and counts-based methods. Circulation 1982; 65: 725-730. 11. Dougherty AH, Naccarelli GV, Gray EL, Hicks CH, Golstein RA: Congestive heart failure with normal systolic function. Am J Cardiol 1984; 54: 778-782. 12. Soufer R, Wohlgelernter D, Vita NA, etal:lntact systolic left ventricular function in clinical congestive heart failure. Am J Cardiol 1985; 55: 1032-1036. 13. Lee DC, Johnson RA. Bingham JB, eta/:Heartfailure in outpatients. Arandomized trial of digoxin versus placebo. N Engl J Med 1982; 306: 699-705. 14. Aronson JK: Digitalis intoxication. Clin Sci 1983: 64: 253-258. 15. Belier GA, SmithTW. Abelmann WH, Hager E, Hood WB: Digitalis intoxicationa prospective clinical study with serum level correlations. N Engl J Med 1971; 284:
989-997. 16. The Captopril-Digoxin
Multicenter Research Group: Comparative effects of therapy with captopril and digoxin in patients with mild to moderate heart failure. JAMA 1988; 259: 539-544. 17. DiBianco R, Shabetai R, Kostuk W, Moran J, Schlant R, Wright R, and the Milrinone Multicenter Trial Group: Oral milrinone and digoxin in heart failure: results of a placebo-controlled, prospective trial of each agent and the combination (abstr). Circulation 1987; 76: IV-256. 18. Ozawa W, Smith D. Craige E: Origin of the third heart sound. I. Studies in dogs. Circulation 1983; 67: 393-398. 19. Ozawa W. Smith D, Craige E: Origin of the third heart sound. II. Studies in humans. Circulation 1983; 67: 399-404. 20. Ishimitsu T, Smith D. Berko B, Craige E: Origin of the third heart sound: comparison of ventricular wall dynamics in hyperdynamic and hypodynamic types. J Am Coll Cardiol 1985; 5: 268-272. 21. Smith TW, Braunwald E. Kelly RA: The management of heart failure. In: Braunwald E, ed. Heart disease. A textbook of cardiovascular medicine. Philadelphia: WB Saunders, 1988; 504-507.
85
Ca/ifornia
PURPOSE:Digoxin is the third most commonly prescribed drug, yet limited information exists about its use in outpatients. Therefore, 242 medical outpatients receiving digoxin at our hospital were studied to evaluate the appropriateness of its use, defined by: (1) current or past supraventricular arrhythmias and/or (2) left ventricular systolic dysfunction (ejection fraction less than 45 percent). PATIENTSANDMETHODS: Charts of242 patients receiving digoxin were obtained. The patients were divided into groups based upon their physician’s stated indication for digoxin therapy. Patients with only a clinical diagnosis of congestive heart failure (CHF) underwent echocardiography or radionuelide angiography to quantify left ventricular systolic function. Those with documented supraventricular arrhythmias and/or those with confirmed left ventricular systolic dysfunction were classified as appropriate candidates for digoxin. RESULTS:Ninety-five percent of patients received digoxin for appropriate indications; 75 percent had confirmed supraventricular arrhythmias (27 percent also had CHF) and 20 percent with normal sinus rhythm had documented systolic dysfunction. However, physicians had difficulty in the clinical assessment of left ventricular function; 18 percent of patients with sinus rhythm and CHF by the Framingham scoring system and 20 percent of those with supraventricular arrhythmias and CHF had preserved systolic function, An S3 was present in 15 percent of patients with preserved ejection fraction and CHF and in 69 percent with low ejection fraction; hypertension was significantly more common in the former group. Noninvasive assessment of systolic function was obtained in 97 percent of patients independent of this study, yet some patients without supraventricular arrhythmias and with documented preservation of systolic function continued to receive the drug. CONCLUSION: Noninvasive assessment of left ventricular function, which appears to have become routine, is of value in the appropriate utilization of digoxin, since clinicians’ assessment of left ventricular function may be inaccurate. Physicians also do not always discontinue digoxin therapy when indicated.
From the Divisions of General Internal Medicine and Cardiology, San Francisco Veterans Administration Medical Center and the University of California, San Francisco, San Francisco, California. Requests for reprints should tration Medical Center, 4150Clement fornia 94121. Manuscript submitted
Street, March
111 Al, San Francisco, Cali18, 1988. and accepted in
D
igoxin is the third most commonly prescribed drug in the United States [l]. There are two major indications for its use: control of the ventricular response in supraventricular tachyarrhythmias and improvement of myocardial performance in congestive heart failure (CHF) secondary to dilated cardiomyopathy. A number of studies have questioned the efficacy of digoxin for patients with CHF and sinus rhythm, but several of these have included subjects without documented left ventricular systolic dysfunction [26]. In recent years, the frequent occurrence of symptoms and signs of CHF due primarily to diastolic dysfunction and episodic CHF due to ischemic left ventricular dysfunction has been recognized [7]. Response to digoxin would not be expected in such patients. To our knowledge, only one previous study has examined the prevalence of inappropriate usage of digoxin in outpatients [8]; 42 percent of patients were found to be receiving the drug for a questionable reason. Our study was designed to answer the following questions: (1) Do internists prescribe digoxin appropriately based on clinical criteria? (2) Of those patients appropriately prescribed digoxin clinically, how many are potentially benefiting from the drug based on demonstration of resting systolic dysfunction? (3) Is noninvasive assessment of left ventricular function indicated for outpatients receiving digoxin for clinical findings of CHF?
PATIENTS AND METHODS Setting The San Francisco Veterans Administration Medical Center (SFVAMC) is a primary teaching hospital of the University of California, San Francisco School of Medicine. All faculty and house staff hold university appointments. The SFVAMC Group Practices (general medical clinics) are staffed by 24 categorical Internal Medicine house staff, six Division of General Internal Medicine faculty, and three nurse practitioners. Patients Three thousand patients are enrolled in the SFVAMC Group Practices. Since the Veterans Administration health care system provides patient medication, it was possible to identify those receiving digoxin. Pharmacy records from September 1, 1986, of all Group Practice patients were reviewed. Charts were available for 242 of 245 patients receiving digoxin; they constitute the sample for this study. Protocol The protocol for this study was approved by the University of California, San Francisco, Human and Environmental Protection Committee. The following data were abstracted from hospital
September
1988
The American
Journal
of Medicine
Volume
85
365
DIGOXIN
USE IN OUTPATIENTS
/ PAPADAKIS
AND MASSIE
ferred for echocardiography or, if the result was technically unsatisfactory, radionuclide angiography, to quantify left ventricular function. The patients were divided into the following three groups based upon their physician’s stated indication for digoxin therapy. GROUP 1. Patients with current or past supraventriculaf arrhythmias receiving digoxin for control of ventricular response. GROUP 2. Patients with normal sinus rhythm receiving digoxin for the treatment of CHF. GROUP 3. Patients with supraventricular arrhythmias and the clinical diagnosis of CHF.
TABLEI Demographic Data and Indication for Digoxin Use in 242 Study Patients Number Male sex Mean age (years) Mean length of digoxin use (years) Group 1 (supraventricular arrhythmias) Group 2 (CHF) Group 3 (CHF and arrhythmias) All values expressed
as f standard
deviation
(percent)
240 (99) 70f 11* 4f5 !Y $ij 65 (27) unless stated
Study Criteria for Appropriate Digoxin Use Patients with documented supraventricular arrhythmias and those with confirmed left ventricular systolic dysfunction (ejection fraction less than 45 percent by echocardiography, radionuclide angiography, or cardiac catheterization) were classified as appropriate candidates for digoxin. Although in part arbitrary, this value represents the lower 95 percent confidence limits for a normal echocardiographic ejection fraction in our laboratory and a liberal value for the degree of left ventricular systolic dysfunction usually associated with clinical CHF [lo-121.
TABLEII Appropriateness of Digoxin Therapy Based on Physician’s Indication for Use Number Prescribed (percent) Group 1 (suoraventricular krhythmias) Group 2 (CHF) Group 3 (CHF and arrhythmias)
Percent with Preserved Ejection Fraction
Percent Inappropriate 0
116 (48) 61(25) 65 (27)
18 0
:i
TABLEIll Comparison of CHF Patients with Decreased and Preserved Ejection Fraction (EF)
SB
Hypertension
Preserved EF
30%
15%’
58%1
Dir$nI,‘,“,EF (n = 102)
90%
69%
30%
I p
and clinic chart review: patient’s age, sex, cardiac history, initiation date, and indication for digoxin therapy. Results of physical examination, electrocardiography, and chest radiography at the time of institution of digoxin therapy were noted as well as findings on the most recent physical examinations. Data from objective assessment of left ventricular function by echocardiography, radionuclide techniques, hemodynamic measurements, and left ventricular angiography were obtained from the charts and respective laboratories. Tests whose results were equivocal were reviewed for a definitive interpretation by an observer unaware of the clinical presentation. Clinical CHF Scoring System For patients who had the diagnosis of CHF, its severity was quantified by the investigators based on the scoring system developed by the Framingham study group [9]. The presence of an SBgallop was also noted. Prospective Studies Patients with only a clinical diagnosis of CHF (with or without supraventricular arrhythmias) were re-
366
September
1988
The American
Journal
of Medicine
Volume
Analysis Proportions were compared with a chi-square test and statistical significance was set at the 0.05 level.
RESULTS Demographic data and indications for digoxin use in the 242 study patients are shown in Table I. Most were older men who had received digoxin for several years, reflecting the population served by the SFVAMC. Seventy-five percent of patients were receiving digoxin for control of supraventricular arrhythmias, 27 percent with and 48 percent without CHF. CHF in normal sinus rhythm was the indication for the drug in the remaining 25 percent of patients. The patients’ health care providers had independently performed noninvasive assessment of left ventricular function in nearly all patients with the clinical diagnosis of CHF. Ninety-seven percent of Groups 2 and 3 patients had a mean of 2.2 tests (range, one to 12). Eighteen percent of Group 2 and 20 percent of Group 3 patients had an ejection fraction greater than 45 percent by noninvasive assessment (Table II). Of those with a preserved ejection fraction, less than one third had CHF by the Framingham clinical scoring system (Table III). The mean ejection fraction was 29 percent in the remainder of Groups 2 and 3 patients, and 90 percent of them had CHF by the Framingham scoring criteria. Of patients with a diagnosis of CHF and a normal ejection fraction on their most recent evaluation, only one had a previous abnormal ejection fraction. An S3 was reported as present in 15 percent of patients with preserved ejection fraction and CHF, compared with the majority (69 percent) of those with low ejection fraction; the sensitivity and specificity of SS for decreased left ventricular ejection fraction was only 69 percent and 85 percent, respectively. The most common accompanying diagnosis in patients with the diagnosis of CHF and preserved ejection fraction was hypertension, which was present in
85
DIGOXIN
58 percent (Table III). Significantly fewer patients (30 percent) with a reduced ejection fraction had a history of hypertension.
COMMENTS Ninety-five percent of medical outpatients at the SFVAMC were receiving digoxin for appropriate indications based upon predetermined criteria derived from the literature. This included 75 percent who had confirmed supraventricular arrhythmias and 20 percent with normal sinus rhythm but with documented left ventricular systolic dysfunction. This percentage compares favorably with that in previous studies by Carlson et al [8] and Lee et al [13], which demonstrated that more than 40 percent of patients were receiving the drug for questionable indications. The patient populations in those studies differed from ours, however. More of our patients had supraventricular arrhythmias as an indication for digoxin therapy than in the study of Carlson et al, and only those with sinus rhythm were evaluated by Lee and colleagues. Our greater number of patients with supraventricular arrhythmias may serve to increase the proportion of patients receiving the drug appropriately. The major problem with inappropriate digoxin use in our study was difficulty in the clinical assessment of left ventricular function; 18 percent of patients receiving digoxin for CHF with sinus rhythm and 20 percent of patients with supraventricular arrhythmias and the diagnosis of CHF had preserved systolic function. This finding is important because patients with sinus rhythm together with preserved systolic function probably do not benefit from digoxin, so this treatment carries unnecessary risk. Those with preserved left ventricular systolic function and supraventricular arrhythmias but with the diagnosis of CHF may have been candidates for alternative or concomitant treatment with verapamil or beta blockers for optimal rate control and other indications, such as hypertension. The efficacy of digitalis as a positive inotropic agent in patients with CHF continues to be debated [14,15]. The issue assumes particular importance because of the potential toxicity of digoxin and the recent availability of alternative effective agents for CHF. In one of the few controlled trials of digoxin use, a beneficial effect of the drug in outpatients was shown, especially in those with a ventricular gallop (Sa) or systolic dysfunction. However, 44 percent of the patients had no appreciable clinical improvement with the drug [13]. Importantly, this study suggests that the subset of patients who clinically benefit from digoxin have chronic systolic dysfunction, rather than episodic symptoms of CHF or diastolic dysfunction. Furthermore, patients with normal left ventricular ejection fraction did not benefit despite their symptoms of CHF. Indeed, many of the studies that have questioned the efficacy of digoxin have included patients with relatively preserved left ventricular systolic function or have not included such measurements. A recently published report comparing digoxin, captopril, and placebo in mild to moderate CHF [16] confirmed the value of digoxin in at least a subset of patients with mild to moderate symptoms and ejection fractions below 40 percent, but did not show an advantage over captopril. A large study comparing digoxin and milrinone in patients with moderate symptoms recently re-
USE IN
OUTPATIENTS / PAPADAKIS AND MASSIE
ported in preliminary form indicated that digoxin was particularly effective in the subset of patients with ejection fractions less than 20 percent [17]. These reports emphasize the importance of appropriately choosing candidates for digoxin, and the requirement for underlying moderate or severe systolic dysfunction for clinical efficacy. Recent studies employing noninvasive techniques have shown that CHF is often associated with relatively preserved systolic function, and that these patients may be clinically indistinguishable from those with severe contractile dysfunction [11,12]. In studies of patients with classic signs and symptoms of CHF referred for radionuclide angiography, approximately 40 percent did not have significantly abnormal systolic function [11,12]. We found that 18 percent of patients receiving digoxin with a diagnosis of CHF had preserved left ventricular ejection fraction. Explanations for this observation include the following: (1) The diagnosis of CHF is incorrect. (2) CHF is caused by predominantly diastolic, rather than systolic, dysfunction. (3) CHF has resolved. (4) Systolic function has normalized with digoxin therapy. If any one of the aforementioned first three reasons is correct, digoxin should be discontinued. Yet despite documentation of preserved ejection fraction in our study, health providers failed to discontinue the drug. Potentially, patients could receive digoxin indefinitely after treatment is instituted. Even if initiation of the drug were begun by another physician, the current physician should reassess whether a continued need exists; if the initial or current indications for the drug are unclear, it should be discontinued. Although digoxin can produce a modest increase in ejection fraction (less than 5 percent), it is rarely of the magnitude to restore the fraction to the normal range [16]. Hence, normal left ventricular function by noninvasive testing in patients with normal sinus rhythm is adequate justification to discontinue digoxin in most patients. One potential exception would be patients with significant mitral regurgitation whose ejection fraction may be inappropriately high in relation to the degree of myocardial dysfunction. The presence of a left ventricular Sa has been considered one of the best predictors of therapeutic response to digoxin, being both very sensitive and specific [13]. Surprisingly, when contrasted to therapeutic response to digoxin, an Ss is much less sensitive but still very specific for decreased left ventricular ejection fraction [13]. Whereas an Sa can occur normally in younger persons with rapid left ventricular filling and in mitral insufficiency, its pathogenesis in left ventricular failure is probably related to the sudden limitation in the filling movement of the left ventricle during diastole [18-201. Yet our study shows that, in clinical practice, detection of an S3 gallop is neither sensitive nor specific for decreased ejection fraction. We interpret these findings as demonstrating that cardiac auscultation by the clinician is not as accurate as that by the investigators, whose observations were verified by phonocardiography or apexcardiography; cardiac auscultation remains a subjective and variable method. Others have noted the presence of an SS in patients with intact systolic function but with the diagnosis of CHF. An Sawas heard in more than 40 percent of older patients with normal ejection fraction and absence of
September
1988
The American
Journal
of Medicine
Volume
85
367
DIGOXIN
USE IN OUTPATIENTS
/ PAPADAKIS
AND MASSIE
hemodynamically significant valvular heart disease, but who met the Framingham criteria for CHF [ll]. Clinical Implications
Noninvasive assessment of left ventricular function appears to have become routine; nearly all patients in this study had received such a test, but physicians sometimes failed to act on the information. On the basis of our demonstration of the inaccuracy of the physical examination for reduced systolic function, we recommend noninvasive assessment of left ventricular function be undertaken in patients diagnosed with congestive heart failure, especially if sinus rhythm is normal. The cost of two-dimensional echocardiography or radionuclide angiography is $331* and $460*, respectively. Although a year’s supply of digoxin is relatively inexpensive ($86*), additional costs include monitoring of therapeutic drug levels ($67.25 per ievel*), physician office visits, and potential drug toxicity. The therapeutic-to-toxic ratio of digitalis is quite narrow, and its toxicity is one of the most prevalent adverse drug reactions encountered in clinical practice [21]. Toxicity may develop in as many as 5 to 15 percent of hospitalized patients receiving digitalis [21]. Equally important, if no impairment of ejection fraction can be demonstrated, digoxin should be withheld or discontinued. * University
of California,
San Francisco
outpatient
charges,
1988.
REFERENCES l.Top 200 drugs of 1987. American Druggist. February 1988; 197: 36-52. 2. Fleg JL, Gottlieb SH, Lakatta EG: Is digoxin really important in treatment of compensated heart failure? Am J Med 1982: 73: 244-250. 3. Gheorghiade M, Belier G: Effects of discontinuing maintenance digoxin therapy in patients with ischemic heart disease and congestive heart failure in sinus rhythm. Am J Cardiol 1983; 51: 1243-1249. 4. Hull SM, Mackintosh A: Discontinuation of maintenance digoxin therapy in gen-
368
September
1988
The American
Journal
of Medicine
Volume
eral practice. Lancet 1977; II: 1054-1055. 5. Johnston GD, McDevitt DG: Is maintenance digoxin necessary in patients with sinus rhythm. Lancet 1979; II: 567-570. 6. Smith Tw: Medical treatment of advanced congestive heart failure; digitalis and diuretics. In: Braunwald E, Moth MB, Watson JT, eds. Congestive heart failure. New York: Grune & Stratton, 1982; 261-278. 7. Parker JO, Ledwich JR, West RO, Case RB: Reversible cardiac failure during angina pectoris: hemodynamic effects of atrial pacihg in coronary artery disease. Circulation 1969; 39: 745-757. 8. Carlson KJ, Lee DC, Goroll AH, Leahy M, Johnson RA: An analysis of physicians’ reasons for prescribing long-term digitalis therapy in outpatients. J Chronic Dis
1985; 38: 733-739. 9. McKee PA, Castelli WP, McNamara PM, Kannel WB: The natural history of congestive heart failure: the Framingham study. N Engl J Med 1971; 285: 1441-
1446. 10. Massie BM. Kramer BL, Gertz EW, Henderson SG: Radionuclide measurement of left ventricular volume: comparison of geometric and counts-based methods. Circulation 1982; 65: 725-730. 11. Dougherty AH, Naccarelli GV, Gray EL, Hicks CH, Golstein RA: Congestive heart failure with normal systolic function. Am J Cardiol 1984; 54: 778-782. 12. Soufer R, Wohlgelernter D, Vita NA, etal:lntact systolic left ventricular function in clinical congestive heart failure. Am J Cardiol 1985; 55: 1032-1036. 13. Lee DC, Johnson RA. Bingham JB, eta/:Heartfailure in outpatients. Arandomized trial of digoxin versus placebo. N Engl J Med 1982; 306: 699-705. 14. Aronson JK: Digitalis intoxication. Clin Sci 1983: 64: 253-258. 15. Belier GA, SmithTW. Abelmann WH, Hager E, Hood WB: Digitalis intoxicationa prospective clinical study with serum level correlations. N Engl J Med 1971; 284:
989-997. 16. The Captopril-Digoxin
Multicenter Research Group: Comparative effects of therapy with captopril and digoxin in patients with mild to moderate heart failure. JAMA 1988; 259: 539-544. 17. DiBianco R, Shabetai R, Kostuk W, Moran J, Schlant R, Wright R, and the Milrinone Multicenter Trial Group: Oral milrinone and digoxin in heart failure: results of a placebo-controlled, prospective trial of each agent and the combination (abstr). Circulation 1987; 76: IV-256. 18. Ozawa W, Smith D. Craige E: Origin of the third heart sound. I. Studies in dogs. Circulation 1983; 67: 393-398. 19. Ozawa W. Smith D, Craige E: Origin of the third heart sound. II. Studies in humans. Circulation 1983; 67: 399-404. 20. Ishimitsu T, Smith D. Berko B, Craige E: Origin of the third heart sound: comparison of ventricular wall dynamics in hyperdynamic and hypodynamic types. J Am Coll Cardiol 1985; 5: 268-272. 21. Smith TW, Braunwald E. Kelly RA: The management of heart failure. In: Braunwald E, ed. Heart disease. A textbook of cardiovascular medicine. Philadelphia: WB Saunders, 1988; 504-507.
85