8 Mycobacterium avium complex and Mycobacterium kansasii


*E-mail: martin.fisher@bsuh.nhs.uk

8.1 Methods

The PubMed database was searched under the following heading: HIV or AIDS and atypical mycobacterial infections, Mycobacterium avium complex or Mycobacterium avium intracellulare and M. kansasii.

8.2 Introduction

Many atypical mycobacteria have been reported to be isolated and/or cause disease in patients with HIV infection. This is typically in the context of very advanced immunosuppression (CD4 counts of <50 cells/μL) and with most patients having disseminated focal disease. The commonest of these infections are M. avium complex (MAC) and M. kansasii.

Since these organisms are frequently commensals from multiple environmental sources, it is important that a clinical decision is made that the organism is considered to be the cause of disease rather than an incidental finding prior to any specific treatment initiation.

With the exception of MAC, there is limited evidence to guide decisions of choice or duration of therapy and expert opinion should be sought from a clinician experienced in mycobacterial disease. Most of the recommendations for the treatment of atypical mycobacteria have been extrapolated from trials in HIV-seronegative individuals. Where an individual is markedly immunosuppressed, some physicians may increase the number of antimycobacterial agents and/or the duration of therapy.

8.3 Mycobacterium avium complex

8.3.1 Background and epidemiology

Mycobacterium avium complex (MAC) organisms are present throughout the environment. Mycobacterium avium is the predominant atypical mycobacterium that affects patients with HIV-1. Infection occurs via the respiratory or gastrointestinal tracts. No particular activity is known to predispose to infection, and person-to-person transmission is not believed to occur.

Prior to the advent of HAART, disseminated MAC was seen in 40% of patients with advanced HIV infection [1], with even higher rates at autopsy [2]. Since the start of the widespread use of HAART, the incidence of MAC has reduced significantly [3] and the prognosis has improved markedly [4], although the clinical picture has changed to include immune reconstitution disease [5] and focal infection.

8.3.2 Presentation

Disseminated MAC infection (DMAC) typically occurs in patients with advanced immunosuppression who have CD4 counts <50 cells/μL. Patients with DMAC frequently have nonspecific symptoms, signs and laboratory abnormalities, which may be attributed incorrectly to HIV progression or other HIV-related illnesses. Patients most commonly report fever, night sweats, fatigue, weight loss, anorexia and diarrhoea. Common signs include hepatomegaly and lymphadenopathy while laboratory abnormalities include anaemia, leukopenia, elevated alkaline phosphatase levels and hypoalbuminaemia. Radiological features include hepatosplenomegaly and intra-abdominal lymphadenopathy, which were demonstrated in one case series using abdominal computed tomography (CT) in 14 of 17 patients with DMAC [6]. More unusual focal manifestations of MAC infection include palatal and gingival ulceration, septic arthritis and osteomyelitis, endophthalmitis, pericarditis, pulmonary and focal lymphadenitis [4,7–10].

8.3.3 Diagnosis

  • Diagnosis of DMAC requires culture in blood or from a bone marrow aspirate or fluid from a normally sterile site or biopsy specimen (category III recommendation).
  • Culture of MAC from sputum or stool in the absence of other proof of DMAC is insufficient diagnostic evidence (category III recommendation).

Definitive diagnosis of DMAC requires culture of the organism from a sterile body site. Isolation of MAC from non-sterile sites (sputum or stools) in the absence of clinical/radiological features suggestive of disseminated infection is insufficient to warrant antimycobacterial therapy. In some situations empirical treatment may be considered pending the results of culture given the time period necessary before cultures can be reliably deemed negative.

Mycobacterial blood culture (standard and liquid) establishes the diagnosis in 86–98% of cases in which disseminated MAC infection is confirmed at autopsy [11,12]. One blood culture identifies 91% of patients with MAC bacteraemia, a second blood culture increases the identification rate to 98% [13]. Therefore, obtaining paired or more than two sequential blood specimens for culture to diagnose MAC bacteraemia is unnecessary [14].

The preferred culture method includes lysis of peripheral blood leukocytes to release intracellular mycobacteria followed by inoculation on to solid media (e.g. Lowenstein–Jensen, Middlebrook 7H11 agar) or into radiometric broth [15]. Using the radiometric detection system, mycobacteraemia can be detected in as little as 6–12 days, whereas 15–40 days are required with solid media. In addition, DNA probes can identify MAC species within 2 h once sufficient mycobacterial growth has occurred in radiometric broth or on solid media [16]. Multiplex PCR have also been shown to provide a low-cost alternative to DNA probe methods for rapid identification of MAC [17].

Biopsies from other normally sterile body sites can prove diagnostic. Stains of biopsy specimens from bone marrow, lymph node or liver may demonstrate acid-fast organisms or granulomata weeks before positive blood culture results are obtained [18,19].

8.3.4 Treatment Treatment regimens for DMAC.

  • Antimycobacterial treatment of DMAC requires combination therapy that should include a macrolide and ethambutol, with or without rifabutin (category Ib recommendation).
  • Addition of rifabutin should be considered in cases with advanced immunosuppression, marked symptoms or an inability to construct an effective HAART regimen (category III recommendation).
  • Treatment failure is usually addressed with a combination of at least two new drugs including fluoroquinolones, amikacin and other agents (category III recommendation).
  • Treatment should be continued until there is an adequate immunological response to HAART or in the absence of this life-long (category III recommendation)

Macrolide-containing regimens are associated with superior clinical outcomes in randomized clinical trials as compared to non-macrolide-containing regimens [20] (category Ib recommendation). Clarithromycin and azithromycin have both demonstrated clinical and microbiological activity in a number of studies; however, macrolide monotherapy is associated with rapid emergence of resistance [21]. Clarithromycin has been studied more extensively than azithromycin and is associated with more rapid clearance of MAC from the blood [22,23]. However, azithromycin has fewer drug interactions and is better tolerated [24]. The dose of clarithromycin should not exceed 500 mg bd as higher doses have been associated with excess mortality [25]. Emergence of macrolide resistance is associated with a return of clinical symptoms and/or increased bacterial counts in some patients [21]. Therefore, addition of at least one further class is recommended.

Ethambutol is the most commonly recommended second drug [25] and its addition to combinations used for MAC treatment reduces the development of macrolide resistance [26,27]. Ethambutol does not interact with currently available antiretroviral agents. A third drug (usually rifabutin) may be included in the regimen. One randomized clinical trial demonstrated that the addition of rifabutin to the combination of clarithromycin and ethambutol improved survival and the chance of complete microbiological response during the study period, though not microbiological clearance at the primary end-point of 12 weeks or relapse rate, while another study showed it reduced emergence of drug resistance [28,29]. Rifabutin dosage should not exceed 300 mg/day (or 450 mg if given with efavirenz or 150 mg three times a week if given with ritonavir) as cases of uveitis have been reported with higher doses, especially when given with clarithromycin [30–32]. It should be noted that many of the benefits of rifabutin were described pre-HAART and the benefits may be more marginal if HAART is administered. Significant drug–drug interactions also exist between rifabutin and antiretroviral agents (protease inhibitors and NNRTIs), which will affect the decision of whether to use rifabutin as a third agent and/or whether to modify the choice of HAART regimen (see Table 8.1). In situations where there is a high risk of short-term mortality the addition of rifabutin should be strongly considered, for instance in persons with:

Table 8.1.   Potential atypical mycobacterial infection treatment and antiretroviral drug interactions
Drug nameInteraction with antiretroviralAction required
  1. Antiretroviral drugs, especially the NNRTIs and boosted PIs, have several important drug–drug interactions. This table lists some examples of drug–drug interactions with antiretroviral agents and drugs used in atypical mycobacterial infections. As data and advice changes frequently, this information should always be interpreted in conjunction with the manufacturer's information (http://www.medicines.org.uk). Other useful web-based reference sources include the Liverpool HIV drug information website (http://www.hiv-druginteractions.org) and the Toronto Clinic (http://www.hivclinic.ca/main/drugs_interact.html).

 ClarithromycinNNRTIs, PIs and boosted PIs can alter clarithromycin levelsSee individual antiretroviral manufacturer's information
Zidovudine levels may be decreasedSeparate dose interval by 1–2 h
Maraviroc likely to be increasedReduce dose of maraviroc (150 mg od)
 RifabutinBoosted PIs increase rifabutin levels
Unboosted PIs increase rifabutin levels
Indinavir and saquinavir should not be used with rifabutin
Reduce rifabutin dose (150 mg three times weekly)
Reduce rifabutin dose to 150 mg od
Efavirenz reducedConsider increasing rifabutin dose (450 mg od)
Etravirine – limited dataUse with caution (standard 300 mg od)
Levels decreased when used in conjunction with a boosted PI
 RifampicinNNRTI levels reducedIncrease dose of efavirenz (800 mg od) depending on patient's weight)
Contraindicated with etravirine and nevirapine (some units increase nevirapine dose)
PI levels significantly reducedNot recommended to be used together. Seek HIV specialist pharmacist advice
Maraviroc levels reducedIncrease maraviroc dose (600 mg bd)
Raltegravir levels reducedConsider increasing dose of raltegravir (800 mg bd)
  • 1Advanced immunosuppression (CD4+ T lymphocyte count <25 cells/μL);
  • 2Markedly symptomatic DMAC features and/or laboratory parameters (as described above);
  • 3A situation where there is an inability to construct an effective HAART regimen.

There are few supporting data for the use of other drugs such as a fluoroquinolones or parenteral amikacin [33]. These should therefore only be considered when rifabutin or other first-line drugs cannot be used because of drug interactions, intolerance or treatment failure. Clofazimine should not be used in the treatment of MAC as it is associated with excessive toxicity and higher mortality rates [34].

In summary, the preferred regimen for disseminated MAC is clarithromycin (500 mg twice daily) or azithromycin (500 mg/day) plus ethambutol (15 mg/kg/day). If rifabutin (usually 300 mg/day) is included in the regimen a dose adjustment is necessary if concurrently administered with a ritonavir-boosted protease inhibitor (150 mg three times a week) or efavirenz (450 mg daily) (seeTable 8.1). Length of treatment for DMAC.

  • Individuals receiving HAART with a virological response and a CD4 count >100 cells/μL for at least 3 months in whom there has been a clinical response to DMAC therapy for at least 3 months can discontinue therapy (category 3 recommendation)

Most studies of the treatment of DMAC were performed in the pre-HAART era. However, there is no doubt that one of the most effective treatments for DMAC is HAART. HAART should be initiated simultaneously or within 1–2 weeks of initiation of antimycobacterial therapy for DMAC disease, based on the experience with a range of opportunistic infections including a small number of cases with MAC [35] (category IV recommendation). If patients are already on HAART at the time of DMAC diagnosis, HAART should be continued and/or adjusted to ensure the viral load is undetectable (<50 copies/mL HIV-1 RNA) (category IV recommendation). Successful initiation of HAART is a key determinant of the duration of DMAC therapy.

The incidence of DMAC has dropped dramatically with the use of HAART. Prior to the HAART era, therapy for DMAC was life-long. It has become clear that immune reconstitution and CD4 cell recovery secondary to HAART enables successful withdrawal of MAC therapy in most cases.

Whilst there are no randomized clinical trial data to strongly recommend duration of MAC therapy after initiation of HAART, prospective non-randomized studies [36,37] and cohort studies [38,39] would suggest DMAC therapy can be safely discontinued in patients responding to HAART. Although studies have differed with regard to the exact conditions that need to be fulfilled before DMAC therapy is stopped, the minimal conditions that should be satisfied are a clinical response to MAC therapy for at least 3 months' duration (asymptomatic with negative blood cultures after 6 weeks incubation) and both of the following:

  • 1virological response to HAART (viral load <50 copies/mL on two consecutive occasions); and
  • 2an immunological response to HAART (confirmed CD4 count >100 cells/μL) on two separate occasions at least 3 months apart.

In the absence of these criteria being met DMAC treatment should be continued lifelong or until these criteria can be met. Treatment failure for DMAC. Patients are considered to have treatment failure if there is no clinical response and mycobacteria are isolated from cultures after 4–8 weeks of MAC treatment to which the patient has been adherent. Drug susceptibility testing is of limited use for agents other than macrolides (category III recommendation). Ethambutol and rifabutin drug susceptibility to MAC has not been correlated to clinical response to therapy although there are data for clarithromycin and azithromycin [40,41]. A new combination of at least two drugs not previously used and to which the isolate should be susceptible should be constructed (category III recommendation) – e.g. rifabutin (if not used previously), ciprofloxacin, levofloxacin, ofloxacin or moxifloxacin [42], linezolid or amikacin. Other second-line agents (such as ethionamide, prothionamide or cycloserine) have been used anecdotally. Many clinicians would continue ethambutol since it facilitates the penetration of other agents into mycobacteria (category IV recommendation).

Immunomodulators, including granulocyte colony-stimulating factor and interferon gamma, can be considered in cases of DMAC treatment failure. They are thought to work by inhibiting intracellular replication or enhancing in vitro intracellular killing of M. avium but there are no comprehensive studies of these agents [43,44]. Treatment of focal MAC. There are no data to guide the type or duration of therapy for focal MAC. However, given that these tend to occur at higher CD4 cell counts and in the presence of effective HAART, most clinicians would recommend a three-drug regimen for a duration of at least 12 and possibly 24 months.

Potential drug interactions may lead to modifications in the HAART and/or antimycobacterial regimen (seeTable 8.1).

8.3.5 Primary prophylaxis

  • Prophylaxis for DMAC with azithromycin 1250 mg weekly can be considered for individuals with CD4 counts <50 cells/μL (category Ib recommendation).
  • Primary prophylaxis can be stopped in the presence of a virological response to HAART (viral load <50 copies per mL) and a CD4 count >50 cells/μL for at least 3 months (category III recommendation).

Randomized clinical trials have demonstrated a benefit of clarithromycin/azithromycin or combinations of rifabutin and azithromycin [45,46] in reducing the incidence of MAC infection in patients with a CD4 count of <100 cells/μL. However, these studies were conducted prior to the introduction of HAART, which has itself resulted in a massive reduction in the incidence of MAC [3]. Furthermore, in one of these studies, where CD4 cell counts at diagnosis of DMAC were provided, it was observed that no cases of DMAC occurred with a CD4 count >50 cells/μL. Thus, lowering the CD4 count at which primary prophylaxis should be considered to <50 cells/μL is recommended in line with many other guidelines. It is therefore recommended that MAC prophylaxis should be considered for individuals with a CD4 count of <50 cells/μL who are either not accepting HAART or who are experiencing HAART failure (category Ib recommendation).

In such individuals, the decision to recommend MAC prophylaxis will need to balance the potential clinical benefits against the additional pill burden, possible added drug-related toxicity, and risk of resistance if undiagnosed DMAC is present (category IV recommendation).

Rifabutin, clarithromycin or azithromycin are acceptable, although azithromycin (1250 mg weekly) is preferred since it has fewer potential drug–drug interactions and is better tolerated (category Ib recommendation). The dose recommended in this guideline differs from the 1200 mg dose traditionally recommended in other guidelines and reflects the size of azithromycin tablets available in the UK.

Primary prophylaxis can be stopped when the patient has a response to HAART (viral load <50 copies per mL) and a CD4 count >50 cells/μL for at least 3 months (category III recommendation). Some physicians prefer to use a cut-off of 100 cells/μL based on evidence from two papers. In these studies, all the patients had CD4 counts >100 cells/μL on stopping prophylaxis, but no cases of DMAC and only two cases of atypical focal MAC were seen [47,48]. No data are available for a >50 cells/μL cut-off. However, owing to the effect of antiviral therapy on MAC, the toxicity of azithromycin seen in prophylaxis studies, and the fact that almost all cases of MAC occur at CD4 counts of less than 50 cells/μL, as evidenced in the Pierce study [46], a cut-off of 50 cells/μL has been considered most appropriate.

8.3.6 Impact of HAART

  • HAART should be commenced within 2 weeks of starting MAC therapy (category IV recommendation).

The incidence of DMAC has dropped dramatically with the use of HAART. HAART should be initiated promptly after diagnosis of MAC and primary and secondary prophylaxis can be discontinued after an initial response to HAART as outlined above.

MAC IRIS can occur as focal disease presenting as regional lymphadenopathy, liver lesions, bone lesions or hypercalcaemia [49–54]. This syndrome is usually self-limiting but can be severe and require adjunctive therapy. There are currently no randomized data to recommend the optimal management strategy. However, the following have been used with anecdotal benefit (category III recommendation) and may be considered in select cases:

  • 1Corticosteroid therapy, with 20–40 mg of oral prednisolone a day for 4–8 weeks has been most frequently used;
  • 2IL-2 and GM-CSF have been used successfully in a small number of patients [55];
  • 3Leukotriene inhibitors have been used in TB-associated IRIS in cases refractory/intolerant to steroids [56,57];
  • 4Repeated fine-gauge needle aspiration of pus, as used with lymphadenitis due to Mycobacterium tuberculosis.

8.4 Mycobacterium kansasii

8.4.1 Background and epidemiology

M. kansasii is the second most common nontuberculous mycobacterium producing disease in patients with HIV infection [58].

Pulmonary disease is seen in over half of patients [58–60], and bacteraemia occurs in fewer than 25% of individuals, although disseminated infection is associated with advanced immunosuppression.

8.4.2 Presentation

Presentation is pulmonary in over half of cases [59–61]. The most typical presenting symptoms/features are fever, cough, focal pulmonary signs on examination and radiological features of pulmonary cavities or infiltrates. Approximately 10% of cases may have mycobacteraemia [62].

8.4.3 Diagnosis

Since M. kansasii may be a commensal organism, diagnosis requires both repeated isolation and a compatible clinical and radiological picture (category IV recommendation).

8.4.4 Treatment

  • Where clinically indicated, treatment is with rifamycin+ethambutol+isoniazid for a minimum of 12 months (category IV recommendation).

The decision to initiate therapy must be clinically based. In patients where M. kansasii is isolated from non-sterile sites (usually sputum) in the absence of clinical and or radiological disease, specific therapy should be withheld. Repeated positive isolates may signify active disease even in the absence of new symptoms.

Therapy should be with a rifamycin such as rifampicin 600 mg od or rifabutin 300 mg od plus ethambutol 15 mg/kg with high-dose isoniazid 300 mg od plus pyridoxine 20 mg od for at least 12 months (category IV recommendation) and possibly for at least 12 months of documented sputum negativity. However, the duration is based on pre-HAART and/or HIV-seronegative extrapolation data (for more details see [63]). There is also experience with the combination of clarithromycin, rifampicin and ethambutol (category IV recommendation).

The treatment regimen for disseminated disease should be the same as for pulmonary disease. Because of the critically important role of rifamycins in the treatment of M. kansasii disease, it is important to construct M. kansasii and antiretroviral treatment regimens that are compatible (see Table 8.1). The recommended regimen for M. kansasii would be rifampicin/rifabutin plus ethambutol plus/minus high-dose isoniazid. An option for treating HIV-seropositive patients who receive an antiretroviral regimen not compatible with rifamycins is to substitute a macrolide or quinolone (e.g. ofloxacin) for the rifamycin. The recommendations for duration of therapy for disseminated M. kansasii disease in patients with HIV are similar to the recommendation for duration of therapy for disseminated MAC infection (above).

8.4.5 Prophylaxis

There is no recommended prophylaxis, and secondary prophylaxis is not indicated for disseminated M. kansasii disease as is the case with M. tuberculosis.

8.4.6 Impact of HAART

There is insufficient data to allow comments on the impact of HAART.