Use of rifabutin for the treatment of a latent tuberculosis infection in a patient after solid organ transplantation


  • The UCSF Committee on Human Research approved the study protocol.

Address reprint requests to Matthew D. Hickey, B.S., School of Medicine, University of California San Francisco, 513 Parnassus Avenue, Room S-245, Box 0454, San Francisco, CA 94143-0454. Telephone: 415-937-1651; E-mail:; Fax 415-502-1320


Latent tuberculosis infection is an important problem for solid organ transplant recipients because of the frequency of its occurrence and its potential for reactivation. Because of the high mortality rate associated with active tuberculosis infections in transplant recipients, guidelines from the American Thoracic Society recommend treatment for latent tuberculosis in this population. However, the choice of treatments is often difficult because liver transplant recipients may be more sensitive to isoniazid hepatotoxicity, and rifampin has significant drug interactions with the calcineurin inhibitors used for immunosuppression. Two prior case reports described success with the use of rifabutin, a rifampin alternative, as part of a multidrug treatment regimen for active tuberculosis in posttransplant patients; however, there is no prior literature describing any experience with rifabutin for the treatment of latent tuberculosis in the posttransplant setting. We present a summary of tacrolimus drug levels and corresponding dose requirements for a single posttransplant patient during the administration of 3 different latent tuberculosis drug regimens: rifampin alone, rifampin plus ketoconazole, and rifabutin. In this patient's case, rifabutin allowed the maintenance of adequate tacrolimus levels, although an approximate 2.5-fold increase in the dose was required. Rifampin alone was associated with inadequate immunosuppressant levels, and rifampin plus ketoconazole was associated with a problematically prolonged QT interval and concerns about inadequate tuberculosis treatment. Liver Transpl 19:457–461, 2013. © 2013 AASLD.


cytochrome P450




latent tuberculosis infection.

Latent tuberculosis infection (LTBI) is a common problem for solid organ transplant recipients. One US center recently reported evidence of an LTBI rate of 6% to 9% in patients waiting to receive liver transplants.[1] Active tuberculosis occurs at varying frequencies, with rates of 0.3% to 1.2% among renal transplant recipients in the United States and rates of 0.9% to 6.5% among nonrenal transplant recipients in the United States and Europe. The average time to diagnosis in this systematic review was 9 months after transplantation, with most cases attributed to immunosuppression-induced reactivation of latent infections.[2] Among pediatric patients, a Spanish center recently reported a cumulative active tuberculosis rate of 500 per 100,000 transplant recipients per year; this rate is approximately 40 to 50 times greater than the regional pediatric incidence over the same period.[3] Active infections are associated with a 31% short-term mortality rate in transplant recipients.[4]

Although this has not been adopted by all centers, guidelines advocate LTBI testing before transplantation.[1, 4-10] Monotherapy with isoniazid (INH) for 9 months is recommended as a first-line therapy, with 4 months of rifampin as an alternative choice. Rifabutin has been suggested as a rifampin alternative, although there are few data to support this recommendation in transplant patients.[8, 10, 11] The choice of therapeutic agents and the timing of therapy are complicated and depend on the degree of urgency of the transplant, the risk of drug toxicity, the drug interactions with immunosuppressant medications, and the transplant type.

Among renal transplant candidates and recipients, INH reduces the incidence of active tuberculosis, is not associated with an increase in hepatotoxicity, and is generally the treatment of choice for LTBI.[12-14] INH therapy for LTBI has been more controversial in liver transplant candidates because of concerns about increased hepatotoxicity.[15] Evidence for this concern has been mixed, with some reporting adequate tolerability[16, 17] and a systematic review reporting that 41% of patients experienced hepatotoxicity requiring the cessation of INH.[2] In liver transplant recipients, a recent meta-analysis found that only 6% treated with INH monotherapy for LTBI required discontinuation of therapy, whereas 1% required emergent liver transplantation for INH-induced acute liver failure.[4] This review also found that INH reduced the incidence of active tuberculosis from 8.3% to 0.0% (P = 0.02). Guidelines remain unclear on the optimal time for initiating INH therapy for LTBI; however, some have recommended that treating physicians consider delaying therapy until graft function stabilization after transplantation (Evidence level BIII).[10]

Rifampin therapy for LTBI in transplant patients is generally recommended only when INH is contraindicated or poorly tolerated.[10] It is associated with significant drug interactions with immunosuppressant medications and can thus increase the risk of acute rejection when it is used after transplantation.[11] Rifampin can also cause hepatotoxicity, although the risk with monotherapy is significantly less than that with INH.[18, 19] The major advantages of rifampin for LTBI in patients for whom drug interactions are not a problem are the shorter duration of therapy and thus the higher rate of treatment completion.[18, 19] Rifabutin has been suggested as an alternative treatment option with fewer cytochrome P450 (CYP450) drug interactions, but this option has not been studied for the treatment of LTBI in the posttransplant setting. Because of these limitations and the paucity of studies providing explicit information on clinical outcomes associated with these drug interactions, heterogeneity in practice persists. Here we describe the pharmacokinetic outcomes of 3 different approaches to the treatment of LTBI in a patient concurrently receiving tacrolimus for immunosuppression after solid organ transplantation.


A 14-year-old boy with chronic renal insufficiency secondary to methylmalonic acidemia underwent simultaneous kidney and split graft liver transplantation from a deceased donor. His methylmalonic acidemia had been diagnosed at 5 days of life, and he was notable for numerous hospitalizations for metabolic decompensation progressing to coma. His preoperative history was also notable for a purified protein derivative finding of 10 to 12 cm with a negative chest X-ray that was consistent with LTBI. He was from Saudi Arabia, a tuberculosis-endemic country, and had received INH for several months before transplantation. However, this medication was stopped at the time of surgery because of a concern about increased sensitivity to the medication's hepatotoxic effects. The decision was made to restart treatment for his latent tuberculosis with rifampin several months after transplantation when his clinical course had stabilized.

His initial posttransplant immunosuppressive regimen consisted of tacrolimus, mycophenolate mofetil, and a methylprednisolone taper. Three months after transplantation, he was started on rifampin (10 mg/kg) for his LTBI because of the stability of his immunosuppressive regimen and the improvement of his hepatic function after some earlier difficulty with total parenteral nutrition cholestasis. Immediately after he started rifampin, his serum tacrolimus concentration dropped to subtherapeutic levels, and his tacrolimus dose requirements increased significantly (Fig. 1). After 10 days of therapy with rifampin, ketoconazole was added in an attempt to boost tacrolimus levels. Despite improved tacrolimus levels and somewhat decreased dose requirements, the patient developed hyperkalemia in the setting of acute renal failure thought to be due to vancomycin toxicity. An electrocardiogram demonstrated prolonged QT intervals at this time. Because of the patient's prior history of ventricular tachycardia, this prompted the discontinuation of all medications associated with prolonged QT intervals, including ketoconazole. Rifampin was simultaneously discontinued because of the inability to maintain adequate tacrolimus levels without ketoconazole. During this time, the patient also developed recurrent low-grade fevers and leukocytosis along with a focal right upper lobe consolidation and bilateral right-greater-than-left pleural effusions. These findings raised concerns about active tuberculosis; however, the workup, which included induced sputum, bronchoalveolar lavage, and thoracentesis of the effusion fluid, did not uncover any evidence of an active infection. Although a multidrug regimen for active tuberculosis was considered during this diagnostic workup, it was ultimately withheld because of concerns about drug interactions. Instead, the patient remained on LTBI therapy with the initiation of rifabutin monotherapy several days after the cessation of rifampin. The consolidation and fevers soon resolved after a 10-day course of meropenem, and this led to a presumptive diagnosis of pneumonia as an explanation for these fevers. Figure 1 charts the patient's tacrolimus doses and levels over time while he was on various combinations of antituberculosis medications.

Figure 1.

Timeline of the tacrolimus doses and serum concentrations during the administration of various therapies for LTBI.

The initial course with rifabutin (5 mg/kg) was notable for its ability to maintain therapeutic tacrolimus levels. Despite this initial improvement in the management of drug interactions, the non-oliguric acute renal failure that had developed during rifampin administration still did not resolve. A renal biopsy sample taken soon after the initiation of rifabutin demonstrated acute tubular necrosis without signs of rejection. This appeared to be temporally associated with a period of very high vancomycin levels. Because of possible vancomycin-induced acute tubular necrosis, tacrolimus was also withheld because of concerns about synergistic renal toxicity. After a 2-week tacrolimus hiatus, another renal biopsy sample was taken because of continued elevations in his creatinine. This biopsy sample showed moderate rejection, and tacrolimus was subsequently restarted along with a short course of methylprednisolone and a prednisone taper. After the resolution of the patient's acute rejection, his tacrolimus levels were maintained at or near therapeutic levels throughout the remainder of his hospitalization, although close monitoring and increased tacrolimus doses were required. There was no clinical or laboratory indication of hepatotoxicity from any of the tuberculosis drugs administered. The patient was discharged 2 months after the initiation of rifabutin therapy with a plan for the completion of the 2 remaining months of a 4-month course as an outpatient back in Saudi Arabia. Table 1 demonstrates his average daily tacrolimus dose requirements and drug levels while he was on several different LTBI regimens. This table excludes tacrolimus levels determined during the 2-week period while he was on rifabutin but tacrolimus was withheld.

Table 1. Average Tacrolimus Levels and Dose Requirements With Several Concurrent LTBI Treatment Regimens Along With P Values From t Test Comparisons of Levels and Doses During Concurrent LTBI Treatments and the Control Period
 Days Tested (n)Average Tacrolimus Level (mcg/L)P Value for Tacrolimus LevelaAverage Tacrolimus Dose (mg/Day)P Value for Tacrolimus Dosea
  1. a

    Versus the control value.

Control386.39 1.06 
Rifampin plus ketoconazole76.770.642.290.001


Although INH with close liver enzyme monitoring would now be the treatment of choice for this patient at our center, at the time of his transplant several years ago, there was more heterogeneity in practice patterns. Some transplant providers felt that treatment with INH carried too much risk of hepatotoxicity, so there was hesitancy in using it for this patient. On the basis of the management of this patient's LTBI, we are able to provide a unique look at the pharmacokinetics of 3 distinct and clinically important drug interactions between rifamycins and calcineurin inhibitors. Although they are not first-line therapies, these treatment options could potentially be considered when INH is not tolerated or drug resistance is present.[9]

The treatment of our patient with rifampin resulted in a sharp drop in the tacrolimus serum concentration to subtherapeutic levels (therapeutic target = 5-10 mcg/L) despite a steep, approximately 3.8-fold increase in the tacrolimus dose. Both the drug levels and the dose changes were statistically significant when they were averaged over 10 days of rifampin administration and compared with the control period when the patient was not on antituberculosis medications (Table 1). The difficulties in using rifampin in this patient mirror the literature describing the suppressive effect that rifampin has on tacrolimus levels as well as the subsequent risk that this poses for acute graft rejection.[11, 20]

The addition of ketoconazole returned the tacrolimus serum concentration to prerifampin levels, and lower doses of tacrolimus were required in comparison with rifampin alone (Table 1). Ketoconazole inhibits CYP450, and this results in significant increases in calcineurin inhibitor levels.[21] However, ketoconazole's CYP450 activity can also reduce rifampin levels, and this causes concerns about subtherapeutic antituberculosis effects.[20] It can also increase concentrations of other medications and result in potential drug toxicity. This may have been the cause of this patient's prolonged QT interval during ketoconazole administration.[22, 23]

During rifabutin administration, the serum tacrolimus concentrations were maintained at a level similar to the control level; however, this required an approximately 2.5-fold tacrolimus dose increase over the baseline. Although this was less than the increase required during rifampin coadministration, both were clinically significant. These observations are supported by pharmacokinetic studies demonstrating that rifabutin has similar but less pronounced CYP450-inducing properties than rifampin.[24] Clinically, rifabutin and rifampin have similar safety profiles, cure rates, and relapse rates when they are used to treat active pulmonary tuberculosis.[25-28] In transplant recipients, 2 case reports have shown success in substituting rifabutin for rifampin[29] or rifampicin[30] as part of a treatment regimen for active tuberculosis. There are no studies comparing the effectiveness of rifabutin to the effectiveness of INH or rifampin for LTBI. The present case provides the only reported experience with rifabutin for the treatment of LTBI in the post–solid organ transplant setting. Despite these limited data, the difficulty encountered in treating tuberculosis with other drugs in posttransplant patients and the similar effectiveness of rifabutin and rifampin have influenced the Infectious Diseases Society of America and other groups to recommend rifabutin as an appropriate substitute for rifampin in settings in which rifampin is poorly tolerated.[8, 11]

Although uncertainties remain about certain aspects of this case, the serum drug levels available from this patient provide a unique look at the drug interactions involved in LTBI treatment in the posttransplant setting. Our experience with this patient also provides the first clinical data reported for the treatment of LTBI with rifabutin in a patient concurrently taking tacrolimus after solid organ transplantation. In this case, rifampin was not tolerated because of substantial drug interactions with tacrolimus. Although the rifabutin therapy was not without challenges, including the need for an increased tacrolimus dose and constant tacrolimus level monitoring, its distinct advantage over rifampin was its ability to maintain therapeutic tacrolimus levels. Ketoconazole plus rifampin also allowed therapeutic tacrolimus levels but complicated LTBI management by introducing greater opportunities for drug interactions. Further prospective trials are needed to demonstrate the improved tolerance of rifabutin when it is coadministered with calcineurin inhibitors; however, these data suggest that this approach may be merited when rifampin and INH are contraindicated or not tolerated.