Clinically important drug interactions with zopiclone,zolpidem and zaleplon

Insomnia, an inability to initiate or maintain sleep, affects approximately one-third of the American population. Conventional benzodiazepines, such as triazolam and midazolam, were the treatment of choice for short-term insomnia for many years but are associated with adverse effects such as rebound insomnia, withdrawal and dependency. The newer hypnosedatives include zolpidem, zaleplon and zopiclone. These agents may be preferred over conventional benzodiazepines to treat short-term insomnia because they may be less likely to cause significant rebound insomnia or tolerance and are as efficacious as the conventional benzodiazepines. This review aims to summarise the published clinical drug interaction studies involving zolpidem, zaleplon and zopiclone. The pharmacokinetic and pharmacodynamic interactions that may be clinically important are highlighted. Clinical trials have studied potential interactions of zaleplon, zolpidem and zopiclone with the following types of drugs: cytochrome P450 (CYP) inducers (rifampicin), CYP inhibitors (azoles, ritonavir and erythromycin), histamine H(2) receptor antagonists (cimetidine and ranitidine), antidepressants, antipsychotics, antagonists of benzodiazepines and drugs causing sedation. Rifampicin significantly induced the metabolism of the newer hypnosedatives and decreased their sedative effects, indicating that a dose increase of these agents may be necessary when they are administered with rifampicin. Ketoconazole, erythromycin and cimetidine inhibited the metabolism of the newer hypnosedatives and enhanced their sedative effects, suggesting that a dose reduction may be required. Addition of ethanol to treatment with the newer hypnosedatives resulted in additive sedative effects without altering the pharmacokinetic parameters of the drugs. Compared with some of the conventional benzodiazepines, fewer clinically important interactions appear to have been reported in the literature with zaleplon, zolpidem and zopiclone. The fact that these drugs are newer to the market and have not been as extensively studied as the conventional benzodiazepines may be the reason for this. Another explanation may be a difference in CYP metabolism. While triazolam and midazolam are biotransformed almost entirely via CYP3A4, the newer hypnosedatives are biotransformed by several CYP isozymes in addition to CYP3A4, resulting in CYP3A4 inhibitors and inducers having a lesser effect on their biotransformation.     

New drugs for insomnia: comparative tolerability of zopiclone, zolpidem and zaleplon.

Insomnia affects 30-35% of people living in developed countries. The impact of insomnia on daytime functioning and its relationship with medical and psychiatric illnesses necessitate early treatment to prevent insomnia becoming persistent and to avoid the development of complications. However, pharmacological strategies must achieve a balance between sedative and adverse effects. In the last 30 years, benzodiazepines have been the preferred drugs for the treatment of insomnia. Benzodiazepines act nonselectively at two central receptor sites, named omega(1) and omega(2), which are located in different areas of the CNS. The sedative action of benzodiazepines is related to omega(1) receptors, whereas omega(2) receptors are responsible for their effects on memory and cognitive functioning. According to their pharmacokinetic profile, benzodiazepines can be classified into three groups: short half-life (<3 hours), medium half-life (8-24 hours) and long half-life (>24 hours). The newer non-benzodiazepine agents zopiclone, zolpidem and zaleplon have a hypnosedative action comparable with that of benzodiazepines, but they display specific pharmacokinetic and pharmacodynamic properties. These three 'Z' agents all share a short plasma half-life and limited duration of action. In addition, these agents are selective compounds that interact preferentially with omega(1) receptors (sedative effect), whereas benzodiazepines also interact with omega(2) receptors (adverse effects on cognitive performance and memory). Zaleplon is characterised by an ultrashort half-life (approximately 1 hour). Zolpidem and zopiclone have longer half-lives (approximately 2.4 and 5 hours, respectively). These properties, together with the low risk of residual effect, may explain the limited negative influences of these agents on daytime performance. Psychomotor tasks and memory capacities appear to be better preserved by non-benzodiazepine agents than by benzodiazepines. When present, cognitive deficits almost exclusively coincide with the peak plasma concentration. In particular, impairment can emerge in the first hours after drug administration, whereas psychomotor and memory tests carried out 7-8 hours later (i.e. in the morning) generally show no relevant alterations. As with benzodiazepines, the three 'Z' non-benzodiazepine agents should be used for a limited period, even in chronic relapsing conditions. Further evaluation is needed of the safety of hypnosedative medications in the long-term management of insomnia.     

Effect of zaleplon on digoxin pharmacokinetics and pharmacodynamics.

The pharmacokinetics and pharmacodynamics of digoxin alone and digoxin plus zaleplon were studied. Healthy, nonsmoking men between 18 and 45 years of age were given a single oral dose of digoxin 0.375 mg daily on days 1 through 9. On days 10 through 14, the subjects received digoxin 0.375 mg plus oral zaleplon 10 mg daily. Blood samples were obtained on days 3, 5, 8, 9, and 14, and serum digoxin concentration data were analyzed by model-independent pharmacokinetic methods. Blood pressure, heart rate, PR interval, and QTc interval were recorded to determine the effect of zaleplon on digoxin pharmacodynamics. A total of 20 men completed the study. Maximum serum digoxin concentration and area under the serum digoxin concentration-versus-time curve from 0 to 24 hours met bioequivalence test criteria. There were no significant differences in QTc or PR interval between days 9 (digoxin alone) and 14 (digoxin plus zaleplon), and there were no clinically important changes from baseline to the study's end in vital signs, physical examination findings, or ECG results for individual subjects. Eighteen percent of the subjects who received digoxin alone and 35% of those who received digoxin plus zaleplon reported one or more adverse effects; all were mild and resolved quickly. Zaleplon had no significant effects on selected pharmacokinetic and pharmacodynamic properties of digoxin.     

Effect of itraconazole on the pharmacokinetics and pharmacodynamics of zopiclone

Objective: We studied the possible interaction between itraconazole, a potent inhibitor of CYP3A, and zopiclone, a short-acting hypnotic. Methods: A double-blind, randomized, two-phase crossover design was used. Ten healthy young subjects received daily either 200 mg itraconazole or placebo for 4 days. On day 4 they ingested a single 7.5-mg oral dose of zopiclone. Plasma concentrations of zopiclone and itraconazole were determined and pharmacodynamic responses were measured up to 17 h. Results: Itraconazole significantly increased the C_max of zopiclone from 49 to 63 ng ⋅ ml⁻¹. The t_1/2 of zopiclone was prolonged from 5.0 to 7.0 h. The AUC(0–∞) of zopiclone was increased from 415 to 719 ng ⋅ ml⁻¹ h by itraconazole. No statistically significant differences were observed in the pharmacodynamic responses between the groups. Conclusion: Itraconazole has a statistically significant pharmacokinetic interaction with zopiclone but this is only of limited clinical importance, at least in young adults. Received: 15 April 1996 /Accepted in revised form: 4 June 1996     

Effect of itraconazole on the pharmacokinetics and pharmacodynamics of zolpidem

Objective: Zolpidem is a short-acting␣imidazopyridine hypnotic which is biotransformed in humans mainly by CYP3A4. Itraconazole strongly interacts with many substrates of CYP3A4 such as midazolam and triazolam. In this study, the effect of itraconazole on the pharmacokinetics and pharmacodynamics of zolpidem was investigated to uncover a possible clinically significant interaction. Methods: In a randomized cross-over study with two phases, ten healthy volunteers took either 200 mg itraconazole or placebo once daily for 4 days. A single oral dose of 10 mg zolpidem was given on day 4. Plasma drug concentrations were measured up to 17 h and effects of zolpidem up to 9 h after the ingestion of zolpidem. Results: Itraconazole had no marked effects on the pharmacokinetics of zolpidem; the total area under the plasma zolpidem concentration–time curve (AUC_0–∞) was 34% larger during the itraconazole phase (759 ng · h · ml⁻¹) than during the placebo phase (567 ng · h · ml⁻¹). Exophoria of the eyes by the Maddox wing test was significantly increased by itraconazole, but the results of the digit symbol substitution test, critical flicker fusion test, postural sway tests and the visual analogue scale tests for subjective drowsiness and overall drug effect did not differ between the phases. Conclusion: The pharmacokinetics and pharmacodynamics of zolpidem were not remarkably affected by itraconazole in healthy volunteers. Therefore, unlike triazolam, for example, zolpidem can be used in normal or nearly normal doses together with itraconazole and probably also with other CYP3A4 inhibitors. Received: 30 September 1997 / Accepted in revised form: 12 January 1998     

Hypnotic dependence: zolpidem and zopiclone too

(1) Loss of efficacy can occur after a few weeks of treatment with zolpidem and zopiclone. This is not a reason to increase the dose. (2) Zolpidem and zopiclone carry a risk of dependence and of potentially severe disorders (e.g. seizures) during withdrawal. (3) The risk-benefit ratio of benzodiazepines and related drugs (zopiclone and zolpidem) remains more favourable than that of other drugs used in insomnia. However, first-line treatment of sleep disorders should not necessarily be drug-based.     

Comparative tolerability, pharmacodynamics, and pharmacokinetics of a metabolite of a quinolizinone hypnotic and zolpidem in healthy subjects.

The objectives of this double-blind, placebo-controlled study were to assess the single dose tolerability, pharmacodynamics, and pharmacokinetics of Ro 41-3290 (5, 10, and 30 mg) and zolpidem (10 mg) in three sequential groups of 10 healthy male subjects. Pharmacodynamic (tracking, attention, and memory test) and pharmacokinetic measurements were conducted over a period of 24 and 50 h, respectively, after drug intake. Ro 41-3290 was well tolerated at all doses as was zolpidem. Performance in both a tracking and a memory search test was affected at 1.5 h after administration of zolpidem, whereas effects had vanished by 8 h. Ro 41-3290 induced moderate, dose-independent effects, which were most pronounced at 4 h after intake. Long-term memory, as assessed by a word learning and recall test, was not clearly affected by any drug. The pharmacokinetics of Ro 41-3290 were dose proportional with an elimination half-life of approximately 8 h. The relatively slow absorption of Ro 41-3290 (t(max) approximately 2.5 h) and the concentration-effect time delay do not make it a good candidate to replace its parent compound Ro 41-3696 as an investigational hypnotic.     

The effect of erythromycin on the pharmacokinetics and pharmacodynamics of zopiclone.

1. The effect of erythromycin on the pharmacokinetics and pharmacodynamics of oral zopiclone, a non-benzodiazepine hypnotic, was investigated in a double-blind, cross-over study. 2. Ten healthy volunteers were given placebo or 500 mg erythromycin orally three times a day for 6 days followed by an oral dose of 7.5 mg zopiclone. 3. Erythromycin increased plasma zopiclone concentration by 4-fold at 0.5 h (P < 0.05) and by 2-fold at 1 h (P < 0.05). There were increases of 3- and 2-fold in the AUC(0,1 h) and AUC(0,2 h) values (P < 0.05). The total AUC of zopiclone increased by 80% (P < 0.05) but the peak concentration by only 40% (P < 0.05). The peak time of zopiclone concentration was reduced from 2 to 1 h (P < 0.001). 4. Significant pharmacodynamic differences between the treatments were observed from 0.5 h to 2 h with respect to saccadic latency and digit symbol substitution tests. 5. The interaction between erythromycin and zopiclone resulted mainly in accelerated absorption which may lead to a faster hypnotic effect in patients.     

Pharmacoepidemiological characterisation of zolpidem and zopiclone usage

Purpose

Zolpidem and zopiclone are two widely used non-benzodiazepine hypnotics whose usage seems to be associated to pharmacodependence. However, to our knowledge, there has as yet been no published epidemiological study which has compared their abuse or dependence potential. We used a pharmacoepidemiological approach to identify and characterise zolpidem and zopiclone users in real life situations.

Methods

Regular users of zolpidem or zopiclone were identified in the database of a French regional health insurance organisation. A latent class analysis (LCA) was used to identify different subgroups of users of these two hypnotics.

Results

The study cohort comprised 25,168 patients who regularly used zolpidem and 21,860 who regularly used zopiclone. The results of the latent class analysis, which enables subgroups with similar patterns of response to be identified, revealed four clinical subtypes of users of zolpidem: non-problematic users, users with associations with hypnotics/anxiolytics or with associated mental disorders, and problematic users. Only three subgroups were identified for zopiclone, and LCA did not discriminate a special class of problematic users for this drug.

Conclusion

Our analysis indicates that there is a subclass of zolpidem user suggestive of abuse; this was not the case for zopiclone. This methodology is very interesting because it allows analysis of databases and determination of a specific signature of drugs potentially leading to abuse or dependence.     

Effect of voriconazole on the pharmacokinetics and pharmacodynamics of zolpidem in healthy subjects

AIMS: To assess the effect of voriconazole on the pharmacokinetics and pharmacodynamics of zolpidem. METHODS: In a randomized cross-over study with two phases, 10 healthy subjects ingested 10 mg of zolpidem with or without oral voriconazole pretreatment. The concentrations of zolpidem were measured in plasma up to 24 h and pharmacodynamic variables were monitored for 12 h. RESULTS: Voriconazole increased the peak plasma concentration of zolpidem by 1.23-fold [P < 0.05; 90% confidence interval (CI) 1.05, 1.45] and the area under the plasma zolpidem concentration-time curve by 1.48-fold (P < 0.001; 90% CI 1.29, 1.74). The time to peak plasma zolpidem concentration was unchanged by voriconazole but the half-life was prolonged from 3.2 to 4.1 h (P < 0.01; 95% CI on the difference 0.27, 1.45). The pharmacodynamics of zolpidem were unaffected by voriconazole. CONCLUSION: Voriconazole caused a moderate increase in exposure to zolpidem in healthy young subjects but no clear pharmacodynamic changes were observed between the groups.     

Comparative pharmacokinetics and pharmacodynamics of lorazepam,alprazolam and diazepam

The contribution of differential absorption-distribution pharmacokinetics to drug activity can be partially determined by comparing simultaneous estimates of drug serum level with pharmacodynamic effects. In the present paper we have contrasted the effects of clinically equipotent doses of lorazepam, alprazolam, and diazepam on the performance of tracking and digit symbol substitution tasks. Eight young males were tested for 12 h after ingesting the drug. The three benzodiazepines and placebo were administered to each subject according to a balanced double-blind Latin square design. A model is presented that describes the relationship between drug concentration and the degree of impairment across time after the final peak effect. Exponential rate constants were determined for each drug using a Marquardt nonlinear fit of the pooled data. Basically, the constants relate offset serum drug values to the impairment curves at a time when serum-brain equilibrium is assumed to have occurred. The values indicate markedly rapid acute tolerance for alprazolam and diazepam but relatively little acute tolerance for lorazepam. Whether these constants reflect adaptation or differential association-dissociation receptor rate constants cannot be determined, but they do highlight the need to consider receptor kinetics as an important factor in benzodiazepine pharmacodynamics.     

Comparative pharmacokinetics and pharmacodynamics of pravastatin and lovastatin

The oral bioavailability of two HMG-CoA reductase inhibitors, pravastatin and lovastatin, was investigated in this randomized, two-way crossover study. Twenty healthy men were randomly assigned to treatment with a 40-mg dose of pravastatin or lovastatin once daily for 1 week; steady state kinetics were assessed after the last dose. After 1 week of washout, each subject received the alternate treatment. Serum specimens were assayed by gas chromatography/mass spectrometry (GC/MS) for intact pravastatin or lovastatin acid and by bioassay for active inhibitor concentration and, after hydrolysis of lactones, for total inhibitor concentration. The systemic bioavailabilities of total (active plus potentially active) inhibitors for the two drugs were different, with the mean AUC value for lovastatin being 50% higher than that of pravastatin (mean +/- SEM AUC0-24 values of 285 +/- 25 and 189 +/- 13 ng-equiv x hr/mL, respectively, P less than .0001). Pravastatin, which is administered as the monosodium salt, is present in the systemic circulation as the open acid; lovastatin, which is administered as the lactone, is present as both open-acid active metabolites (62%) and closed-ring lactone metabolites (38%), which are potentially active. Based on mean AUC values, pravastatin accounted for 75% of the active inhibitors from a pravastatin dose. Lovastatin acid accounted for just 25% of the active inhibitors from a lovastatin dose, with the remainder due to other active metabolites. Significant decreases from baseline in total and low-density lipoprotein (LDL) cholesterol were observed during the first treatment leg for both pravastatin and lovastatin.(ABSTRACT TRUNCATED AT 250 WORDS)     

Comparison of zolpidem and zaleplon exposures in Texas, 1998-2004

Zolpidem and zaleplon are used for the treatment of insomnia. The objective of this study was to compare the patterns of zolpidem and zaleplon exposures reported to Texas poison control centers during 1998-2004. There were 5842 total reported zolpidem exposures, of which 2918 (50%) were isolated exposures, and 467 total reported zaleplon exposures, of which 201 (43%) were isolated exposures. Zolpidem patients were 62% male and 67% adult. Zaleplon patients were 67% male and 34% adult. The exposure was intentional in 62% of zolpidem and 58% of zaleplon exposures. The exposure occurred at the patient's own residence in 94% of zolpidem and 97% of zaleplon exposures. Management occurred outside of a health care facility for 29% of zolpidem and 32% of zaleplon exposures. The medical outcome involved no symptoms due to exposure for 29% of zolpidem and 44% of zaleplon exposures, a statistically significant difference. Although many of the most frequently reported adverse clinical effects for the two drugs were similar (drowsiness, slurred speech, hallucinations, ataxia, tachycardia, dizziness, confusion, vomiting), the proportion of exposures with a given adverse clinical effect was generally lower for zaleplon. Thus, although zolpidem and zaleplon exposures were generally similar with respect to patient gender and age, exposure reason and site, and management site, zaleplon exposures were less likely to result in minor medical outcomes or manifest as adverse clinical effects.     

Comparative pharmacokinetics and pharmacodynamics of cardiac glycosides.

1 The pharmacokinetics and pharmacodynamics of ouabain, digoxinn and beta-methyl digoxin (medigoxin) have been investigated in a crossover study in four normal healthy volunteers. 2 Pharmacokinetics were studied using [3H]-labelled glycosides and the shortening of the left ventricular ejection time (LVET) was used as a measure of the effect of the drugs. A graded exercise protocol was used to correct for the effects of heart rate on LVET. 3 In three of the four subjects, both digoxin and beta-methyl digoxin produced a shortening in the LVET, but no such change could be detected with ouabain in any of the four subjects. 4 There was a good linear correlation between the shortening of the LVET and the amounts of digoxin or beta-methyl digoxin present in the body tissues. 5 One subject who showed no drug-related LVET shortening had greatly enhanced clearances of all three drugs studied.     

The antinociceptive effect of zolpidem and zopiclone in mice

Zolpidem and zopiclone are two of a newer hypno-sedative class of drugs, the "Z compounds". Their use for the treatment of short-term insomnia has been expanding constantly during the last two decades. The "Z compounds" are considered to cause less significant rebound insomnia or tolerance than the conventional hypnotic benzodiazepines. Their possible antinociceptive effect and interaction with the opioid system has not been studied yet. Our results demonstrate a significant difference between the antinociceptive properties of zopiclone and zolpidem when injected s.c. in the hotplate analgesic assay in mice. Zopiclone induced a weak, dose-dependent antinociceptive effect, antagonized only by the alpha2-adrenergic receptor antagonist yohimbine. Zolpidem induced a weak, biphasic dose-dependent antinociceptive effect, antagonized primarily by the non-selective opioid antagonist naloxone and by yohimbine. The weak antinociceptive effect of both drugs, evident only at very high doses (far beyond those used clinically to induce sleep), implies no clinical use for zopiclone or zolpidem in the management of pain. However, the possible interaction of zolpidem with the opioid system should be further investigated (in behavioral models, which do not overlap with the acute-pain antinociception model we used), both for possible side effects in special populations (i.e. elderly) and for possible drug-drug interactions, in order to minimize possible hazards and maximize clinical beneficial effects of its use for sleep.     

The effect of zolpidem and zopiclone on memory.

The effect of 10 mg of zolpidem (ZLP) on memory function was evaluated in healthy male adults using word recall test, passage recall test, and Sternberg's memory scanning task. This study was carried out as a double-blind cross-over study with 7.5 mg of zopiclone (ZPC) and a placebo. No difference was noted between the active drugs and placebo in the number of words recalled from the word list presented before administration. No evidence of retrograde amnesia was suggested. However, encoding ability was slightly affected as indicated by a decrease in the number of words recalled from the list presented after administration. Slight impairment of a delayed recall was noted for both of the active drugs, but the effect disappeared the next morning. In the memory scanning task, ZLP prolonged a non-specific component of reaction time 1.5 h after administration, but the effect disappeared after 12.5 h. ZPC did not prolong the reaction time. The two active drugs showed no specific effects on either memory scanning or response-selection stage in short-term memory at any time. The findings suggest that residual effects did not reach clinical significance at the standard dosage, although the active drugs slightly affected encoding ability and retention soon after administration.     

Comparative pharmacodynamics and pharmacokinetics of conventional and long-acting propranolol

This investigation was conducted to compare the pharmacokinetic and pharmacodynamic effects of single and multiple doses of conventional propranolol and long-acting propranolol in healthy human volunteers. Two double-blind, randomized, double-crossover, Latin square studies were carried out. One study evaluated long-acting propranolol 160 mg/d, conventional propranolol 40 mg qid, or placebo for seven days in 24 men. The other study compared long-acting propranolol 80 mg/d, conventional propranolol 20 mg qid, or placebo for seven days in 27 men. At specific times after the administration, blood samples were obtained, and heart rate and blood pressure were measured; exercise tests were done both on the first day and at steady state (day 7). In both studies, the area under the plasma propranolol concentration-time curve and the peak concentration were significantly less (P less than .0001) after the administration of long-acting propranolol compared with conventional propranolol on both day 1 and day 7; in addition, the elimination half-life was longer after administration of the long-acting preparation (9 hr) compared with that following the conventional dosage form (4 hr). Both conventional and long-acting propranolol significantly decreased the exercise heart rate at each of the selected time points (P less than .05) compared with placebo. Reduction in exercise heart rate was greater with conventional propranolol than with the long-acting formulation, but the differences were not statistically significant, when exercise was performed only at trough levels of the conventional drug. The decreases in exercise heart rate were correlated with plasma propranolol concentrations.     

Comparative pharmacokinetics and pharmacodynamics of the newer fluoroquinolone antibacterials

A number of new fluoroquinolone antibacterials have been released for clinical use in recent years. These new agents exhibit enhanced activity against Gram-positive organisms while retaining much of the Gram-negative activity of the earlier agents within the same class. The pharmacokinetics of most of these agents are well described including serum pharmacokinetics, tissue and fluid distribution, and pharmacokinetics in renal and hepatic disease. When compared with earlier agents within this class (i.e. ciprofloxacin), the newer agents retain the wide distribution characteristics; however, they exhibit a more prolonged elimination, which, in part, supports single daily administration for these agents. Based on their predominant renal elimination, dosage adjustment is necessary in the presence of renal disease for ciprofloxacin, levofloxacin, gatifloxacin and sitafloxacin. Drug interactions, particularly with multivalent cations (calcium/aluminium-containing antacids and iron products), remain a problem for the newer agents, resulting in reduced absorption requiring separate administration times to maximise bioavailability. However, the newer agents do not appear to interfere significantly with the cytochrome P450 system, thus minimising the potential for interactions with other drugs metabolised by this system. The pharmacodynamic properties of the fluoroquinolones have been well described. The bactericidal activity is maximised when the ratios of peak plasma drug concentration (Cmax): minimum inhibitory concentrations (MIC) or area under the concentration-time curve (AUC): MIC exceed specific threshold values. Knowledge of the pharmacodynamic relationships allows for appropriate drug selection and enables design of dosage regimens to maximise the bactericidal activity. Therapeutic drug monitoring of the fluoroquinolones may provide a means of optimising the dosage regimen in certain clinical situations (that is, meningitis and hospitalised pneumonias) with the goals of achieving a more predictable therapeutic response and minimising the potential for the development of resistance.     

Comparative pharmacokinetics and pharmacodynamics of the rifamycin antibacterials.

The rifamycin antibacterials, rifampicin (rifampin), rifabutin and rifapentine, are uniquely potent in the treatment of patients with tuberculosis and chronic staphylococcal infections. Absorption is variably affected by food; the maximal concentration of rifampicin is decreased by food, whereas rifapentine absorption is increased in the presence of food. The rifamycins are well-known inducers of enzyme systems involved in the metabolism of many drugs, most notably those metabolised by cytochrome P450 (CYP) 3A. The relative potency of the rifamycins as CYP3A inducers is rifampin > rifapentine > rifabutin; rifabutin is also a CYP3A substrate. The antituberculosis activity of rifampicin is decreased by a modest dose reduction from 600 to 450mg. This somewhat surprising finding may be due to the binding of rifampicin to serum proteins, limiting free, active concentrations of the drug. However, increasing the administration interval (after the first 2 to 8 weeks of therapy) has little effect on the sterilising activity of rifampicin, suggesting that relatively brief exposures to a critical concentration of rifampicin are sufficient to kill intermittently metabolising mycobacterial populations. The high protein binding of rifapentine (97%) may explain the suboptimal efficacy of the currently recommended dose of this drug. The toxicity of rifampicin is related to dose and administration interval, with increasing rates of presumed hypersensitivity with higher doses combined with administration frequency of once weekly or less. Rifabutin toxicity is related to dose and concomitant use of CYP3A inhibitors. The rifamycins illustrate the complexity of predicting the pharmacodynamics of treatment of an intracellular pathogen with the capacity for dormancy.