The Role of Genetically Determined Polymorphic Drug Metabolism in the Beta-Blockade Produced by Propafenone

Fourteen healthy volunteers (eight men and six women; mean age [±SD], 32±7 years) took part in this study after providing written informed consent. In 13 of the subjects, the phenotype for the hydroxylation of debrisoquin was determined from the debrisoquin metabolic ratio.¹⁴ Biotransformation of the antihypertensive agent debrisoquin to its 4-hydroxy metabolite is catalyzed by P-450_dbl; subjects with the poor-metabolizer phenotype were defined as those who had a urinary ratio of debrisoquin to 4-hydroxy debrisoquin that was greater than 12.8 after the administration of 10 mg of debrisoquin. In the remaining subject, the P-450_dbl phenotype was established by determining the ratio of 5-hydroxy propafenone to propafenone in serum.^{10 , 11}

Study Protocol

The first five days of the study consisted of a placebo phase in all patients (one tablet every eight hours). Patients were blinded to all treatments (including placebo). On day 5, an isoproterenol-sensitivity test was performed according to the method of Cleaveland et al.¹⁵ with the subjects in a fasting state before the administration of the morning dose. The isoproterenol dose that increased the heart rate by 25 beats per minute (CD₂₅) was determined by linear regression analysis of the steep portion of the logarithm of the dose–response curve for isoproterenol. On day 6, eight hours after the last dose, the subjects performed a treadmill exercise test according to the standard protocol of Bruce et al.¹⁶ to determine their maximal exercise heart rate (HR_max) and underwent pulmonary-function tests to determine the forced expiratory volume in one second (FEV₁). Each subject underwent a treadmill evaluation with use of the Bruce protocol before entry into the study. Thereafter, all repeat evaluations during the placebo phase and drug treatment were terminated at the same stage and duration of exercise.

Subjects were then assigned to receive one of the three propafenone regimens for five days (150, 225, or 300 mg every eight hours) in a double-blind fashion. For reasons of safety, the first dosage was always either 150 or 225 mg every eight hours (assigned randomly). Isoproterenol-sensitivity tests, treadmill tests, and pulmonary-function tests were performed on days 5 and 6 as described for the placebo phase. In addition, trough blood samples were obtained eight hours after the last propafenone dose on the morning of day 6 of each treatment period and the concentrations of propafenone, 5-hydroxy propafenone, and N-desalkyl propafenone were measured by high-performance liquid chromatography.¹⁷ After a 48-hour washout period, the next regimen was then begun. Except as described below, each subject received placebo and all three propafenone regimens.

Affinity of Beta₂ Receptors for Propafenone and Its Metabolites

The binding of propafenone and its metabolites to beta₂ receptors on human lymphocytes was also assessed. Blood was drawn from five healthy, drug-free, nonsmoking male volunteers (age, 22 to 52 years) with EDTA used as the anticoagulant. Mononuclear-leukocyte membranes were prepared as previously described.¹⁸ Competition-binding studies were performed with the radioligand [¹²⁵I]iodopindolol (8.14×10¹³ Bq [2200 Ci] per millimole; New England Nuclear, Boston) in the presence of increasing concentrations of propafenone, 5-hydroxy propafenone, or N-desalkyl propafenone (10⁻¹⁰ to 10⁻⁴ mol per liter, in duplicate). The data for propafenone alone have been reported previously.³ Curves describing the competition for binding [¹²⁵I]iodopindolol were analyzed with a four-parameter logistic equation¹⁹ that provided estimates of the concentration of the competitor that occupied 50 percent of the specific binding sites (IC₅₀). A constant (K_i) describing the affinity of the receptor for propafenone or its metabolites was determined with the program LIGAND.²⁰

Statistical Analysis

Repeated-measures analysis of variance was used to investigate the dose-dependency of beta-blocking actions. The relation between the plasma concentration of propafenone (and its metabolites) and beta-blocking effects was analyzed initially by simple linear regression and then by multiple linear regression. This analysis used one data point per subject, obtained by comparing effects during the first propafenone regimen (150 or 225 mg every eight hours) to those with placebo, since not all subjects completed the full protocol (see below). The measures of beta-blockade used in this analysis were the reduction in the peak exercise heart rate (HR_max during propafenone treatment — HR_max during placebo) and the isoproterenol dose ratio (CD₂₅ during propafenone treatment/CD₂₅ during placebo).¹⁵ In addition, the relation between effect and concentration was analyzed for each subject individually (with data from all regimens used), and the concentrations that produced a 10 percent reduction in the maximal exercise heart rate and a dose ratio of 4 were determined by linear regression. Data on subjects with either of the two phenotypes were compared by two-tailed Student's t-tests (unpaired). All results are expressed as means ±SD.

Nine subjects had the extensive-metabolizer phenotype, and five had the poor-metabolizer phenotype. All the subjects with the extensive-metabolizer phenotype completed the protocol. Severe nausea prevented two of the five subjects with the poor-metabolizer phenotype from completing the protocol; one did not receive the full five days of 225 or 300 mg of propafenone every eight hours, and the other did not receive the full period of treatment with 300 mg every eight hours (plasma propafenone level, 4.93 to 5.40 μmol per liter [1682 to 1842 ng per milliliter]).

Figure 1. Figure 1. Beta-Blocking Effects of Propafenone as a Function of the Dose in Subjects with the Extensive-Metabolizer Phenotype (Open Bars) and the Poor-Metabolizer Phenotype (Hatched Bars).

The upper panel shows the effect of propafenone on CD₂₅ (the dose of isoproterenol necessary to increase the heart rate by 25 beats per minute). The lower panel shows the effect of propafenone on the maximal heart rate during treadmill exercise (HR_max). The changes in these indexes of beta-blockade produced by propafenone were significant in both groups (P<0.0001 by analysis of variance). The asterisks indicate a significant difference (P<0.05) between the subjects with extensive metabolism and those with poor metabolism.

Table 1. Table 1. Steady-State Plasma Concentrations of Propafenone and Its Metabolites, According to Metabolizer Phenotype.*

Figure 2. Figure 2. Beta-Blocking Effects of Propafenone as a Function of the Plasma Propafenone Concentration in Subjects with the Extensive-Metabolizer Phenotype (○) and the Poor-Metabolizer Phenotype (●).

Data obtained during the initial propafenone regimen are compared with those obtained during the placebo phase. The upper panel shows the relation between the plasma propafenone concentration and the isoproterenol dose ratio, the ratio of the dose of isoproterenol necessary to raise the heart rate by 25 beats per minute (CD₂₅) during drug treatment to that during the placebo phase, an index of beta-blockade.¹⁵ The lower panel shows the relation between the plasma propafenone concentration and the percent decrease in the maximal heart rate (HR_max) during a treadmill exercise test. To convert values for propafenone to nanograms per milliliter, multiply by 341.

Propafenone produced significant dose-related beta-blockade in subjects with both phenotypes (P<0.0001 by analysis of variance) (Fig. 1). The extent of the increases in CD₂₅ and the reductions in the maximal exercise heart rate were greater in subjects with the poor-metabolizer phenotype, especially at the lower dosages. As Table 1 indicates, these subjects also had significantly higher trough plasma propafenone concentrations, particularly at the lower dosages. The correlations between the plasma concentration of propafenone and its beta-blocking effects (Fig. 2) support the concept that the accumulation of propafenone in plasma accounted for the variable extent of beta-blockade. No changes in FEV1 were found in these normal subjects.

Although the plasma concentrations of propafenone far exceeded those of its metabolites (Table 1), the possibility that the metabolites contributed to the beta-blocking action of propafenone was also evaluated. Simple linear regression analysis indicated significant correlations between the plasma N-desalkyl propafenone level (but not the plasma 5-hydroxy propafenone level) and beta-blocking effects. However, the plasma N-desalkyl propafenone level also correlated highly with plasma concentrations of the parent drug (r = 0.78, P<0.001), and multiple linear regression incorporating concentrations of both propafenone and N-desalkyl propafenone did not indicate that there was a contribution by the N-desalkyl metabolite. Moreover, the concentrations of propafenone estimated to produce a fixed degree of beta-blockade in individual subjects were similar in those with either of the two phenotypes, again suggesting that the metabolites had no major role in mediating beta-blockade. The concentrations estimated to produce a 10 percent reduction in the maximal exercise heart rate were 3.12±1.50 μmol per liter (1065±513 ng per milliliter; range, 0.80 to 5.13 μmol per liter [273 to 1750 ng per milliliter]) in subjects with the extensive-metabolizer phenotype (n = 6) and 3.30±1.10 μmol per liter (1126±377 ng per milliliter; range, 2.09 to 4.76 μmol per liter [715 to 1625 ng per milliliter]; P not significant) in those with the poor-metabolizer phenotype (n = 3). The concentrations estimated to produce a dose ratio of 4 were 3.45±3.54 μmol per liter (1177±1209 ng per milliliter; range, 0.98 to 11.20 μmol per liter [333 to 3824 ng per milliliter]) in subjects with the extensive-metabolizer phenotype (n = 7) and 4.42±0.83 μmol per liter (1509±283 ng per milliliter; range, 3.77 to 5.59 μmol per liter [1288 to 1907 ng per milliliter]; P not significant) in those with the poor-metabolizer phenotype (n = 3).

Figure 3. Figure 3. Displacement of [¹²⁵I]Iodopindolol from β₂ Adrenoceptors on Human Lymphocytes by Propafenone (○), N-Desalkyl Propafenone (■), and 5-Hydroxy Propafenone (△).

The shape of all three curves was monophasic, and the slopes were not significantly different from unity. In five experiments, the displacement curves for 5-hydroxy propafenone (IC₅₀ = 478±154 nmol per liter; K_i = 181±61 nmol per liter) and for N-desalkyl propafenone (IC₅₀ = 90±38 nmol per liter; K_i = 37±15 nmol per liter) were shifted to the right of that for propafenone (IC₅₀ = 32±19 nmol per liter; K_i = 12±4 nmol per liter).

Further evidence that the beta-blocking actions of propafenone were mediated primarily by the parent drug was provided by the in vitro studies (Fig. 3). Propafenone (IC₅₀ = 32±19 nmol per liter) had almost three times greater affinity for the beta receptor than the N-desalkyl metabolite (IC₅₀ = 90±38 nmol per liter) and approximately 15 times greater affinity than the 5-hydroxy metabolite (IC₅₀ = 478±154 nmol per liter).

Propafenone, which has the ability to block cardiac sodium channels, is useful in the treatment of ventricular and atrial arrhythmias. Clinical reports of variable beta-blockade during propafenone therapy have been at variance with the drug's readily demonstrated and consistent effects in vitro. In this study we have shown that a major determinant of this apparent variation in beta-blockade is genetically determined polymorphic propafenone metabolism. Subjects with the extensive-metabolizer phenotype had significantly less beta-blockade than did those with the poor-metabolizer phenotype. Deficient 5-hydroxylation in subjects with the poor-metabolizer phenotype was associated with elevated plasma propafenone and N-desalkyl propafenone levels, which presumably indicated shunting of propafenone metabolism to N-dealkylation in this group. However, the data on the plasma concentrations (Table 1), in combination with the relative affinities of the parent drug and its metabolites at the beta-receptor (Fig. 3), strongly indicate that the accumulation of propafenone itself, and not its metabolites, accounts for the beta-blockade produced by the drug.

Our group has previously demonstrated that side effects are significantly more frequent when trough plasma propafenone concentrations exceed 2.64 μmol per liter (900 ng per milliliter). Such elevated concentrations occur more frequently in subjects with the poor-metabolizer phenotype¹⁰; in this group, the plasma propafenone level increased in linear fashion as a function of the dose. In contrast, in subjects with the extensive-metabolizer phenotype enrolled in this and previous studies,^{10 , 21} plasma propafenone concentration rose disproportionately as the dose increased, suggesting that the P-450_dbl pathway is saturable. These data on disposition agree with the dose–response data (Fig. 1), which show a greater difference between subjects with the two phenotypes at the lower dosages of propafenone than at the higher dosages. Thus, the subjects with the poor-metabolizer phenotype had significantly more beta-blockade at lower dosages, whereas the difference between phenotypes was not statistically significant at the higher dosages. The dosage range used in this study was within that required for the long-term suppression of arrhythmia. In contrast to the concentration-dependence of beta-blockade, we have found that the prolongation of the QRS complex, an in vivo marker of sodium-channel blockade, is more prominent at any given plasma concentration of propafenone in subjects with the extensive-metabolizer phenotype than in those with the poor-metabolizer phenotype¹⁰; this finding indicates that metabolites contribute to sodium-channel blockade but not to beta-blockade. The beta-blocking effect produced by the higher plasma concentrations of propafenone, the reduction of the maximal exercise heart rate by 10 percent, is equivalent to that produced by approximately 0.15 μmol per liter (40 ng per milliliter) of propranolol,²² or 10 to 20 mg of propranolol every eight hours. In contrast to the case with propafenone, however, this concentration of propranolol is much lower than that required to demonstrate "membrane-stabilizing" effects in humans.²³

Genetically determined polymorphic drug metabolism can contribute to variations in individual response to some drugs. For example, the lupus erythematosus syndrome is more common among persons with slow acetylation of procainamide than among those with rapid acetylation.²⁴ More recently, genetically determined defects in drug oxidation have been associated with abnormal drug disposition and may explain some side effects seen at usual drug dosages.¹³ In 7 percent of normal subjects, the biotransformation of debrisoquin to 4-hydroxy debrisoquin is deficient; in these subjects, who have the poor-metabolizer phenotype, the usual doses of debrisoquin are associated with decreased clearance and an increased incidence of symptomatic hypotension.²⁵ Similarly, the alpha-hydroxylation of metoprolol is also determined by P-450_dbl activity. In subjects with the poor-metabolizer phenotype, the plasma metoprolol levels were elevated and beta-blocking action enhanced.²⁶ The consequences of such defective biotransformation of drugs obviously depend on the relative potency of the parent drug and any metabolites whose formation may be perturbed, as well as on the presence of any drugs or disease states that alter drug metabolism or protein binding. For example, in contrast to metoprolol and debrisoquin, which are biotransformed to less active metabolites, the antiarrhythmic agent encainide undergoes P-450_dbl-dependent oxidation to O-desmethyl encainide, which is a more potent sodium-channel blocker by an order of magnitude than the parent drug.²⁷ In subjects with defective P-450_dbl activity, the biotransformation of encainide to O-desmethyl encainide is severely impaired, and the parent drug accumulates in plasma. We have shown that prolongation of the QRS complex, an index of sodium-channel blockade, is significantly greater among subjects with the extensive-metabolizer phenotype who receive single doses of encainide than among those with poor metabolism; moreover, the extent of prolongation of the QRS complex in subjects with the extensive-metabolizer phenotype was markedly decreased when low doses of quinidine, a potent inhibitor of P-450_dbl, were administered along with encainide.²⁸

The 5-hydroxylation of propafenone produces a metabolite that has sodium-channel–blocking potency equal to or greater than that of the parent drug.^{7 , 8} In fact, with propafenone, unlike encainide, the extent of prolongation of the QRS complex during treatment was not altered by the inhibition of P-450_dbl by low-dose quinidine.¹¹ However, our results indicate that the concentration of the parent drug alone determines the extent of a different pharmacologic action, beta-blockade. Beta-blockade during propafenone therapy appears to be important in the genesis of side effects; in addition, beta-blockade may contribute to the suppression of arrhythmias in some patients. Hence, this important drug action, which appears to occur sporadically, is actually predictable, largely on the basis of a well-described genetic variation in the metabolism of the drug.