Comparison of withdrawal phenomena after propranolol, metoprolol, and pindolol

Comparison propranolol,

of withdrawal phenomena metoprolol, and pindolol

after

Three groups of hypertensive patients were studied after they had received one of the three pharmacologically different beta blockers for at least 1 month: pindolol, 10 mg twice a day (n = 7), propranolol median dose, 80 mg three times a day (n = Q), or metoprolol, 150 mg (twice a day (n = 8). After abrupt withdrawal of drug and replacement with placebo, we measured the following on day 0 and approximately every 2 days up to 3 weeks; beta-adrenergic sensitivity (8AS) by the chronotropic dose of isoproterenol to increase heart rate by 25 bpm (CD,,), resting heart rate and blood pressure, and symptoms. All values are medians. On day 0, beta blockade was evident by increased CDs5 values of 618 pg for pindolol, 57 pQ for propranolol, and 10 pg for metoprolol as compared to 2.8, 2.4, and 3.0 fig, respectively, at days 14 to 21, which were considered the ultimate baseline. After pindolol on day 0, the CDZ5 slowly decreased to, but never below, baseline by days 8 to 21. In contrast, after propranolol on day 0, the CD25 decreased significantly two- to fivefold below baseline from days 4 to 14 (tBAS) and after metoprolol two- to threefold below baseline from days 2 to 8. After pindolol, heart rate and blood pressure gradually returned to, but not above, ultimate baseline. In contrast, during the period of tBAS there was a significant overshoot of heart rate in eight of eight patients after metoprolol (day 0 = 61, peak = 88, baseline = 74) but not after propranolol, while a significant overshoot of blood pressure occurred in six of nine patients after propranolol (day 0 = 140/87, peak = 157195, baseline = 140189) but not after metoprolol. Withdrawal symptoms of headache, palpitations, and tremor occurred in one of seven patients after pindolol, six of nine after propranolol, and three of eight after metoprolol. The degree and duration of beta blockade appeared related to drug potency. Withdrawal phenomena occurred after propranolol and metoprolol but not after pindolol. (AM HEART J 104:473, 1982.)

Robert

E. Rangno,

M.D., and Serge Langlois,

M.D.,*

A withdrawal syndrome after abrupt cessation of the beta-adrenergic blocker propranolol is gaining progressive recognition.6s 14.16,3s It has been observed mainly in patients with coronary artery disease, manifesting as worsening of angina, myocardial infarction, atrioventricular arrhythmias, and sudden death.? In hypertensive patients, abrupt withdrawal of beta-blocking drugs has resulted in symptoms of malaise, headache, tremor, sweating, anxiety, and palpitations. 22,34These initial reports led to several prospective studies that attempted to determine the incidence, time course, and mechanism of the phenomena.$ The most viable theory, tested by our group and others, is an adaptive increase in cardiac beta-adrenergic sensitivity> 30,36perhaps as a result of increased receptor number.‘*15 The majority of case reports and formal studies are relative to From the Division Research Institute. Reprint Cedar

requests: Ave., Room

*Recipient tSee

of Clinical

R. E. Rangno, 1131, Montreal,

of a Canadian

references

$See references

Pharmacology,

Heart

Montreal

General

M.D., Montreal General Quebec, Canada H3G Foundation

Hospital

Hospital, lA4.

1650

FeIlowship.

4, 26, 27, 29, 31, 33. 7, 10, 20, 21, 23, 25, 28, 30, 35, 36.

0002-8703/82/080473

+ 06$00.60/O

@ 1982 The

C. V. Mosby

Co.

Montreal,

Quebec, Canada

withdrawal from propranolol, but other beta blockers may also be culpable. Indeed a few reports suggest withdrawal features after metoprolol,22* 23 practolol,32 oxpreno101,32*40 and atenolol,18 but, to our knowledge, not after pindolol (Visken). It is difficult to determine whether any difference in frequency or severity of the withdrawal phenomena is due to limited experience with other beta blockers or to real differences in their pharmacology. We have completed detailed studies in three groups of hypertensive patients and have shown differences after withdrawal from propranolol (nonselective), metopro101 (selective), and pindolol (nonselective with intrinsic sympathomimetic activity). This presentation is limited to a comparison of four measurements made in those three studies, namely, cardiac betaadrenergic sensitivity, resting blood pressure, resting heart rate, and symptoms. METHODS Subjects. Ethical considerations precluded studies in patients with angina; therefore the study population consisted of patients treated for mild to moderate essential hypertension who were free of historical or EGG evidence of coronary artery disease. Table I shows the 473

474

Rangno

and Langlois

Amerlcsn

1

DAYS

after

August. 1992 Heart Journal

WITHDRAWAL

Fig. 1. Median isoproterenol CD,, values (Fg) after abrupt withdrawal of propranolol, metoprolol, and pindolol. The dashedareasrepresent the period of increasedbeta-adrenergic sensitivity after propranolol and metoprolol. After propranolol, sensitivity on day 6 was significantly lessthan day 20 (p < 0.05).

1. Patient characteristics and dose of beta blocker

Table

Age (yr)

No. of Propranolol Metoprolol Pindolol *Daily

doses were divided

patients

Median

Range

Dose fmglday)

9 8 10

48 52 51

(26-79) (42-82) (28-64)

240 300 20

into three times a day for propranolol

and twice a day for metoprolol

patient characteristics and drug doses. Some patients participated in more than one study. All patients were on therapy for at least 1 month before withdrawal Seven patients receiving propranolol continued to take a thiaxide diuretic throughout the study, while others received no other antihypertensive medication. Design. Patients were studied initially on day 0, 12 hours after their last dose of beta blocker, and serially thereafter at the sametime of the day at approximately 2-day intervals for 2 to 3 weeks. Patients refrained from tobacco and caffeine beveragesfor at least 12 hours before each study. Measurements. On eachstudy day, the sameprocedure wasperformed. A brief history wastaken with the useof a checklist to elicit complaints possibly related to withdraw-

Postwithdrawal placebo No Yes Yes

and pindolol.

al. A 19-gauge butterfly needle was inserted into an antecubital vein. After 30 minutes of bed rest in a quiet room, blood pressure was measured three times by a sphygmomanometer.The heart rate changesproduced by small incremental dosesof isoproterenol weremeasuredto obtain an index of cardiac beta-adrenergic sensitivity, according to the method of Cleaveland et a1.8Before each doseof isoproterenol, baselineheart rate was obtained by measuringthe two shortest RR intervals on a resting ECG strip. Isoproterenol, 200 Mg/ml, was used, with 20 and 2 pg/ml solutions obtained by serial 1:lO dilutions. Each bolus doseof isoproterenol was administered by a rapid intravenous push and followed by a rapid infusion of 20 to 30 cc of 5% dextrose solution. The logarithm of each isoproterenol dose was plotted against the resulting

Volume

104

Number

2, Part II

Table

Post-beta-blockade

II. Symptoms after withdrawal of beta blockers Propranolol (n = 9)

Headache Tremulousness Anxiety Chest pain Total

Metoprolol (n = 8)

140

100

1 1 0 0

1 1 0 0

6

3

2

phenomena

160

Pindolol (n = 10)

3 1 1 1

withdrawal

475

PROPRANOLOL

120

80 MtTOPROLOL

180

change in heart rate, to obtain a log dose-responsecurve. A minimum of four points was obtained and plotted by linear least-squares regression. The isoproterenol dose required to produce an increasein heart rate of 25 bpm was obtained from this line of best fit. This dosewill be referred to as the isoproterenol CD,, (chronotropic dose 25). The median CD,, on days 14 to 20 after withdrawal was consideredthe “baseline CDQ5”reflecting beta-adrenergic sensitivity off beta blocker. A decrease of CD,, below baseline reflects a withdrawal-induced increase in cardiac beta-adrenergic sensitivity. The mediansof heart rate and blood pressure on days 14 to 20 were termed “baseline heart rate” and “baseline blood pressure.” Analysis of data. All data for the groups are expressed asmedian and range. Changesin measurementsover time were analyzed by the Friedman nonparametric analysisof variance with multiple comparison and one-tailed probability.” Statistical analysis was performed within each of the three drug groups but not between the groups. RESULTS lsoproterenol CD,, values. The comparative changes in CD,, values after withdrawal from propranolol, metoprolol, and pindolol are shown in Fig. 1. Comparable and reproducible stable baseline CD,, values were achieved by days 14 to 20. The values were 2.4 pg (range 1.4 to 7.6) after propranoloL2.6 pg (range 0.9 to 5.1) after metoprolol, and 3.0 (range 1.6 to 16.7) after pindolol. This contrasts with initial CD,, values of 57 pug (range 4.6 to 355), 8.7 fig (range 3.1 to 24), and 181 pg (range 16.0 to 800) measured 12 hours after the last doses of proprano101, metoprolol, and pindolol, respectively. Not only these initial values but also the subsequent course of CD,, values differed notably between drugs. After propranolol, a period of increased beta-adrenergic sensitivity extended from days 3 to 13. This was manifested by a median CD,, value below baseline during this interval, with all nine subjects having values two- to fivefold below baseline (i.e., increased sensitivity) somewhere between days 5 and 9 @J< 0.05). A similar pattern was observed after withdrawal of metoprolol, which was marked by an earlier onset and shorter duration of increased betaadrenergic sensitivity. Median CD,, values were below baseline between days 2 and 8, with individual

160 140 120 100 80 PINDOLOL

80f 0 t

, 2

, 4

, 6

, 8

t 10

I 14

I 20

DAYS after WITHDRAWAL

Fig. 2. Median systolic and diastolic blood pressures after abrupt withdrawal of propranolol, metoprolol, and pindolol. After propranolol, systolic pressureon day 6 was significantly greater than on day 20 (p < 0.05).

values two- to threefold

subjects however, pindolol, ultimate the CD,, or 20.

below baseline for the eight at some point during the same interval; the differences were not significant. After the median CD,, value fell gradually to the baseline, and in none of the 10 subjects was values on days 0 to 10 lower than on days 14

Blood pressure. Comparative changes in resting blood pressure after drug withdrawal are shown in Fig. 2. After propranolol, six of the nine patients experienced a significant overshoot of both systolic and diastolic pressures; the median of these six subjects was 135/85 initially, 154/98 at peak, and 139/86 at ultimate baseline. Pressures for the entire propranolol group were 140/87 (range 97 to 159/74 to 100) initially, 149/95 (range 128 to 180/84 to 107) on day 6, and 140/89 (range 112 to 180/79 to 120) at baseline. After metoprolol, systolic and diastolic

pressure rose progressively from 149/90 (range 131

to 187/74 to 108) on day 0 to 178/107 (range 147 to 206/91 to 125) on day 14. There was no overshoot

476

Rangno

and Langlois

American

Auguet, 1982 Heart Journal

pindolol (Table II). These were of variable duration, but did not last for more than 4 days except in two patients who had progressive headache concomitant with a rise in their blood pressure after withdrawal of metoprolol or pindolol. DISCUSSION

PINDOLOL

70f

0 t

I 2

I 4 DAYS

I 6

I 8

I 10

I 14

I 20

after WITHDRAWAL

3. Median heart rates after abrupt withdrawal of propranolol, metoprolol, and pindolol. After pindolol, heart rates on days 2, 4, and 6 were significantly greater than days 14 to 20 (p < 0.05).

Fig.

except for one subject, on day 8 (systolic and diastolic). The same progressive rise was observed after pindolol from 144/90 (range 130 to 184/70 to 116) on day 0 to 161/102 (range 140 to 190/80 to 122) on day 20. Six of the 10 patients had a peak of systolic pressure on day 8, varying from 6 to 10 mm Hg above their ultimate baseline; these differences were not significant. Heart rate. Comparative changes in resting heart rate after drug withdrawal are shown in Fig. 3. Two days after propranolol, heart rate had returned to the baseline value of 78 bpm (range 60 to 105), from an initial value of 55 bpm (range 48 to 70) on day 0. After metoprolol, patients had a comparable heart rate of 61 (range 54 to 79) on day 0, but all of them had an overshoot during the period of increased beta-adrenergic sensitivity (median 7 bpm above ultimate baseline, range 3 to 18). The occurrence of these peaks on different days tended to minimize the differences between individual days, so that these differences did not reach statistical significance. Pindolol did not produce the same bradycardia as propranolol and metoprolol, the median initial heart rate being 79 bpm (range 74 to 86). After pindolol withdrawal, an overshoot of heart rate occurred in 9 of 10 patients between days 2 and day 6; it varied from one to 18 bpm (median 10) and was significant (p< 0.05). Symptoms. Clinical symptoms suggestive of withdrawal were described by six patients after propran0101, by three after metoprolol, and by two after

We have previously shown that abrupt withdrawal of chronically administered propranolol is followed by transient increased sensitivity to the chronotropic effects of isoproterenol.30 A similar phenomenon occurred after metoprolol, but the time course of increased beta-adrenergic sensitivity was of earlier onset, shorter duration, and less magnitude than after propranolol. This less pronounced withdrawal phenomenon may be explained by the somewhat lower potency of metoprolol, or perhaps is related to its property of beta-l cardioselectivity. The relatively rapid onset of the increased beta-adrenergic sensitivity after propranolol and metoprolol is consistent with their plasma half-lives of 3 to 6 hours.” The absence of increased cardiac beta-adrenergic sensitivity after abrupt cessation of pindolol contrasts with the results obtained with the other two beta blockers. None of the ten patients who received pindolol had increased sensitivity at any time during the study period and most patients showed a persistent decrease in sensitivity up to 10 days after pindolol was discontinued. This is inconsistent with the plasma half-life of pindolol(3 to 4 hours), which is comparable to that for metoprolol and proprano101.” The cardiovascular and antianginal~ effects of pindolol have been shown to persist much longer than what is expected from plasma half-life and may represent persistence of this more avidly bound drug at the tissue-receptor site. A second plausible explanation for the lack of increased sensitivity after withdrawal of pindolol may relate to its property of intrinsic sympathomimetic activity (ISA), which may down regulate beta-adrenergic receptor responsiveness. Several theories have been proposed concerning the mechanisms of increased beta-adrenergic sensitivity after chronic beta blockade.* The most attractive concept pertains to the interaction between beta-adrenergic receptors and an agonist or antagonist substance. Glaubiger and Lefkowitz15 have shown an increase in beta-receptor concentrations in the hearts of rata chronically treated with propranolol. Similar findings have been recently reported by Aarons et al.’ in

*See

references

1, 6, 7, 14-16,

20, 23, 26.28,

30, 35, 36, 39.

Volume Number

104 2, Part II

human lymphocytes. Conversely, there is evidence that chronic administration of an adrenergic agonist, such as terbutaline, depresses the number of beta receptors on human polymorphonuclear leukocytes13 and various animal tissues.42 Almost teleologically, then, beta agonists down regulate the number of receptors on which they act, and beta antagonists up regulate the number of receptors. The partial agonist effect of pindolol may well counterbalance the antagonist effect on up regulation of the beta receptors. The slow recovery of the down-regulated receptor response may explain the absence of increased sensitivity after withdrawal and the slow return to a normal state of sensitivity. This hypothesis could be supported by measurement of beta receptors and target cell responsiveness after therapy with pindolol in both humans and animals. The overshoot in heart rate or blood pressure and the symptoms as part of the withdrawal features after propranolol, metoprolol, and pindolol were statistically significant in some instances but we believed they were of potential clinical significance even if they occurred in only one patient. Because the patients were resting during our observations, both the frequency and magnitude of the changes could have been underestimated. Some authors have suggested that increased adrenergic sensitivity will only be seen clinically when the sympathetic drive is elevated, as for example by assuming a standing position, exercising, or using a vasodilator.24r38 In addition, our relatively healthy patients without measurable cardiac disease probably had a higher threshold to possible life-threatening consequences of these changes. The clinical implications of these withdrawal features are quite different in patients with ischemic heart disease or arrhythmia. It has been shown that in a given patient, the onset of angina tends to occur at the same degree of myocardial oxygen demand, of which systolic pressure and heart rate are important determinants.37 For ethical reasons, we limited our study to a hypertensive population, but the results can be extrapolated to all patients with overt cardiac disease and to hypertensive patients who may have ischemic heart disease masked by their beta-blocker therapy. The long duration of increased sensitivity, 2 to 8 days after metoprolol and 3 to 13 days after propranolol, suggests that these two drugs should not be rapidly discontinued, whenever possible. We have shown that tapering of propranolol in less than 9 days does not prevent withdrawal phenomena,31*36 while tapering over 2 weeks or 8 10~ dose over 2 weeks prevents the increase in beta-adrenergic sensitivity and most of the other withdrawal fea-

Post-beta-blockade

withdrawal

phenomena

477

tures.12,36 For pindolol, our results show a slow spontaneous return of beta-adrenergic sensitivity to a normal state but should not be interpreted as the definitive demonstration that abrupt withdrawal is innocuous. This should await further confirmation by controlled trials in patients with angina. REFERENCES

RD, Nies AS, Gal J, Hegstrand LR, Molinoff PB: 1. Aarons Elevation of 8-adrenergic receptor density in human lymphocytes after propranolol administration. J Clin Invest 65:949, 1980. WH: Experiments with pindolol (Visken) in healthy 2. Aellig volunteers. Med J Aust 2(suppl):8, 1977. of the duration of action of pindolol 3. Aellig WH: Comparison and slow release oxprenolol in healthy volunteers. Eur J Clin Pharmacol 14:167, 1978. EL, Coltart J, Wettach GE, Harrison DC: Coro4. Alderman nary artery syndromes after sudden propranolol withdrawal. Ann Intern Med 81:625, 1974. 5. Bertil Olsson S, Varnauskas E: Duration of beta-receptors blockade after oral administration of LB-46. Eur J Clin Pharmacol 5:214, 1973. withdrawal (editorial). Lancet 2:592, 1975. 6. Beta-blockade H, Lewis RP, Kates RE, Dalamangas G: Hyper7. Boudoulas sensitivity to adrenergic stimulation after propranolol withdrawal in normal subjects. Ann Intern Med 87:433, 1977. 8. Cleaveland CR, Rango RE, Shand DG: A standardized isoproterenol sensitivity test: The effects of sinus arrhythmias, atropine and propranolol. Arch Int Med 130:47, 1972. 9. Dayer P, Belant L, Gorgia A, Eberlin JL, Fabre J: The duration of pindolol and its relation to plasma levels. J Clin Pharmacol 4:87, 1979. 10. Faulkner SL, Hopkins JT, Boerth RC, Young JL, Jellett L, Nies AS, Bender HW, Shand DG: Time required for complete recovery from chronic propranolol therapy. N Engl J Med 389:605, 1973. 11. Frishman W: Clinical pharmacology of the new beta-adrenergic-blocking drugs. Part 1. Pharmacodynamic and pharmacokinetic properties. AM HEART J 97:663, 1979. 12. Frishman W, Kostis J, Strom J, Hossler N, Elkayam V, Goldner S, Silverman R, Davis R, Weinstein J, Sonnenblick E: Clinical pharmacology of the new beta-adrenergic-blocking drugs. Part 6. A comparison of pindolol and propranolol in treatment of patients with angina pectoris. The role of sympathomimetic activity. AM HEART J 98526, 1979. 13. Galant SP, Duriseti L, Underwood S, Insel PA: Decreased beta-adrenergic receptors on human polymorphonuclear leukocytes after adrenergic therapy. N Engl J Med 298:933, 1978. 14. Gerber JG, Nies AS: Abrupt withdrawal of cardiovascular drugs. N Engl J Med 301:1234, 1979. 15. Glaubiger G, Lefkowitz RJ: Elevated beta-adrenergic receptor number after chronic propranolol treatment. Biochem Biophys Res Commun 78:720, 1977. 16. Harrison DC, Alderman EL: Discontinuation of propranolol therapy-cause of rebound angina pectoris and acute coronary events. Chest 69:1, 1976. 17. Hollander M, Wolfex DA: Non-parametic statistical methods. New York, 1973, John Wiley & Sons, Inc, p 138. 18. Jackson G, Schwartz J, Kates RE, Harrison DC: Physiologic basis for beta blockade rebound: Atenolol vs propranolol. Clin Res 27:444, 1979. 19. Jennings CL, Bobik A, Fagan ET, Korner PI: Pindolol pharmacokinetics in relation to time course of inhibition of exercise tachycardia. Br J Clin Pharmacol 7:245, 1979. 20. Kristensen B, Steiness E, Weeke J: The pathogenesis of propranolol withdrawal syndrome in essential hypertension. Clin Sci 57(suppl 5):4179, 1979.

478

Rangno

and Langlois

21. Leaman DM, Levenson LW, Shiroff RA, Babb JD, DeJoseph RL, Hayes AH, Zelis R: Persistence of biologic activity after disappearance of propranolol from the serum. J Thorac Cardiovasc Surg 72:67, 1976. 22. Lederballe Pedersen 0: Comaarison of metroarolol and hydrochlorothiazide as antihypertensive agents. Eur J Clin Pharmacol 10:381, 1976. Pedersen 0, Mikkelsen E, Lanng Nielsen J, 23. Lederballe Christensen NJ: Abrupt withdrawal of beta-blocking agents in patients with arterial hypertension. Effects on blood pressure, heart rate and plasma catecholamines and prolactin. Eur J Clin Pharmacol l&215, 1979. 24. Lewis MJ, Ross PJ, Henderson AH: Rebound effect after stopping beta-blockers. Br Med J 2:606, 1979. 25. Malling TJB, Dollery CT: Changes in blood pressure, heart rate and plasma noradrenaline concentration after sudden withdrawal of uropranolol. Br Med J 2:366. 1979. 26. Miller RR, Olson HG, Amsterdam EA, Mason DT: Propran0101 withdrawal rebound phenomenon. Exacerbation of coronary events after abrupt cessation of antianginal therapy. N Engl J Med 293:416, 1975. 27. Miller RR, Olson HG, Amsterdam EA, Mason DT: Propran0101 withdrawal rebound phenomenon. Exacerbation of coronary events after abrupt cessation of antianginal therapy. N Engl J Med 293:416, 1975. LD: Hemodynamic and metabolic 28. Myers JH, Horwitz response after abrupt withdrawal of long-term propranolol. Circulation 58:196, 1978. 29. Myers MG, Wisenberg G: Sudden withdrawal of propranolol in patients with angina pectoris. Chest 71:24, 1977. 30. Nattel S, Rangno RF, Van Leon G: Mechanism of proprano101 withdrawal phenomena. Circulation 58:1158, 1978.

American

August. IQ82 Heart Journal

31. Nattel SQ, Shanks J, Rangno RE: Propranolol withdrawal. Ann Intern Med 89:288, 1978. 32. Nellen MS: Trial of two beta-blockade drugs, Ciba 39089 and ICI 50172 in angna pectoris. S Afr Med J 43(suppl):155, 1969. 33. Nigam PD, Malhatra AS: Prolonged action of pindolol. Br Med J 1:742, 1973. 34. O’Brien ET, MacKinnon J: Propranolol and polythiazide in treatment of hypertension. Br Heart J 34:1042, 1972. 35. Pantano JA, Lee YC: Abrupt propranolol withdrawal and myocardial contractility-A study of effects in normal man. Arch Int Med 138:867, 1976. 36. Rangno RE, Nattel S, Lutterodt A: Prevention of propranolol withdrawal phenomena by gradual dose reduction. Circulation (In press.) 37. Robinson BF: Relation of heart rate and systolic blood pressure to the onset of pain in angina pectoris. Circulation 35:1073, 1967. 38. Ross PJ. Lewis MJ. Henderson AH: Beta-blockers withdrawal syndrome. Lancet 1:875, 1978. 39. Shand DG, Wood AJJ: Propranolol withdrawal syndromewhy? Circulation 58:202, 1978. 40. Wilson DF, Watson OF, Peely JS, Turner AS: Trasicor in angina pectoris: A double-blind trial. Br Med J 2:155, 1969. 41. Wilson M, Organ G, Morgan T: The effect on blood pressure of p-adrenoceptor blocking drugs administered once daily and their duration of action when therapy is ceased. Br J Clin Pharmacol 3:857, 1976. 42. Wolfe BB, Harden TK, Molinoff PB: In vitro study of @-adrenergic receptors. Ann Rev Pharmacol Toxic01 17:575, 1977.