Rifabutin to the Rescue?

(See the Major Article by Karau et al, on pages 1498–504.)

In a 1970 review, Stanley Plotkin [1] listed rifampin as the most interesting new antibiotic, specifically noting its in vitro activity against methicillin-resistant staphylococci and its good oral bioavailability. Not mentioned, but soon recognized, were 2 inherent problems with rifampin for treatment of staphylococcal infections: (1) emergence of resistance and therapeutic failure when rifampin is used as a single agent [2] and (2) its potential for pharmacological interactions with many other drugs [3, 4].

Administration of rifampin in combination with another active agent was found to prevent emergence of resistance [2, 5]. Subsequent laboratory and clinical research elucidated other therapeutically useful properties of rifampin, such as good tissue penetration, activity against intracellular bacteria, and bactericidal activity, including activity against bacteria within biofilms [6]. Biofilms, which form on implanted devices during staphylococcal infection, create a metabolic and microstructural environment that renders bacteria phenotypically resistant both to host defenses and to most antibiotics and, thus, biofilm infections are difficult to eradicate. The activity of rifampin against bacteria within biofilms is a strong rationale for current recommendations for its use in combination therapy to treat device-related staphylococcal infections, eg, prosthetic joint infections, spinal implant infections, prosthetic-valve endocarditis, cerebrospinal fluid shunt infections.

One problem that has not been satisfactorily addressed is drug interactions, which are routinely encountered and often preclude the use of rifampin. Rifampin has a long rap sheet of drug interactions. Drugs.com lists 486 drugs that can interact with rifampin, of which 191 are considered to be major interactions. Affected drugs have in common being metabolized by cytochrome P-450 enzymes, of which rifampin is a strong inducer. A partial list includes anticonvulsants, antiarrhythmics, antiestrogens, antipsychotics, oral anticoagulants, antifungals, antibacterials, antiretrovirals, barbiturates, β-blockers, benzodiazepines and related drugs, calcium channel blockers, corticosteroids, cardiac glycosides preparations, clofibrate, oral contraceptives, estrogens, oral hypoglycemic agents, immunosuppressive agents, levothyroxine, losartan, narcotic analgesics, methadone, progestins, selective 5-HT3 receptor antagonists, statins, theophylline, thiazolidinediones, and tricyclic antidepressants [7]. Drug intolerance and toxicity further limit the ability to use rifampin combination therapies.

Rifamycins (eg, rifampin, rifabutin, rifapentine, and refalazil) all have the same mechanism of action. They inhibit bacterial RNA polymerase by binding to the β-subunit. Resistance to any one rifamycin, almost always due to β-subunit mutations that reduce binding, typically confers cross-resistance to other rifamycins. One would expect, therefore, that these agents might be therapeutically comparable and even interchangeable in terms of spectrum and antibacterial activity. Indeed, studies published over 15 years ago demonstrated that in vitro susceptibilities of staphylococci to rifampin, rifapentine, rifabutin are virtually identical [8, 9], a finding confirmed in more recent studies [10–12].

Where rifamycins do substantively differ from each other is in their pharmacokinetics, adverse effects, and drug interactions [12]. Rifabutin is a less broad and less potent inducer of cytochrome P450 enzymes, CTP3A primarily, as well as a substrate of this enzyme, is not an inducer of P-glycoprotein, and is less hepatotoxic, allowing it to be used in circumstances where rifampin cannot be used. Rifapentine, although less well characterized, is more similar to rifampin. It is better absorbed orally than rifabutin, achieves higher serum concentrations, and can be dosed twice weekly. None of these differences would matter for either drug as an alternative to rifampin in treatment of staphylococcal device-related infections, unless they are similarly active against bacteria in vivo, including those within biofilms. The study by Karau et al [13]. in this issue of The Journal of Infectious Diseases, seeks to answer this key question.

The activities of rifampin, rifabutin, and rifapentine, alone and in combination with vancomycin, were compared in a methicillin-resistant Staphylococcus aureus (MRSA) chronic foreign-body infection model of osteomyelitis in rats. Treatment was administered for 15 days, with serum concentrations approximating those achievable in humans. Although the 3 rifamycins behaved in a similar fashion, overall, rifapentine as a single agent or in combination outperformed rifampin and rifabutin. Bacterial counts of bone and wire were significantly lower in rats treated with rifapentine or rifapentine plus vancomycin, compared with rats given no treatment or vancomycin as a single agent. Rifampin was ineffective as a single agent, but in combination with vancomycin it was more effective in clearing organisms from both bone and wire than no treatment or vancomycin as a single agent.

Rifabutin alone was also ineffective. Rifabutin in combination with vancomycin was more effective than no treatment or vancomycin alone in clearing organisms from bone, but not from wire, falling just short of the false discovery–corrected P value threshold of .05. The vancomycin-rifapentine combination was more effective in clearing organisms from bone than vancomycin-rifabutin, but not significantly different in clearing organisms from bone or wire than vancomycin-rifampin. Rifamycin resistance emerged in only 1 of 24 rats treated with rifamycin as a single agent, probably because bacterial burdens in vivo were below the 10⁸–colony-forming unit threshold at which resistant mutants occur. Thus, it is not possible to assess in this model whether one rifamycin is better than another in preventing emergence of resistance on therapy, or whehter there are differences in efficacy that might be evident at higher bacterial loads.

Extrapolating data from animal models to clinical practice is difficult. After all, these experiments tested only a single strain of MRSA, and not a coagulase-negative strain of staphylococci, and only a single companion antibiotic, vancomycin. How faithfully the model reproduces a chronic device-related infection is difficult to assess with (1) a 4-week interval between infection and treatment, (2) a relatively low bacterial burden, and (3) an end point that relies on quantitative culture at 15 days rather than relapse after a prolonged period off antibiotics. On the other hand, the fact that vancomycin alone was ineffective, whereas the rifamycin combinations were effective in clearing organisms, is evidence of biofilm formation in vivo.

Despite the limitations, the results do suggest that either rifabutin or rifapentine could be an alternative to rifampin in combination therapy of staphylococcal device-related infections. Which is the better alternative, rifabutin or rifapentine? The latter appeared to be more active, but the data are too limited to make much of this difference. With respect to avoiding drug-drug interactions, rifabutin has the better profile compared with rifapentine. A case series of 10 patients successfully treated with rifabutin combinations for S. aureus hardware, graft, or prosthetic device infections without recurrence [14] is the first direct clinical evidence in support of rifabutin as an alternative to rifampin. A randomized controlled trial is warranted to define the efficacy and advantages of rifabutin compared with rifampin in combination therapy for device-related infections, so that clinicians will not have to rely on personal experience, anecdotal reports, and data from observational studies, as was the case for decades with rifampin.

Note

Potential conflicts of interest. Author certifies no potential conflicts of interest. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References