Comprehensive, up-to-date information on HIV/AIDS treatment, prevention, and policy from the University of California San Francisco

Mycobacterium avium complex (MAC) consists of several related species of mycobacterium that are ubiquitous in the environment. MAC rarely causes disease in individuals with a normal immune system. In patients with AIDS, however, it is one of the most common serious opportunistic infections (OIs).(1,2) Among HIV-infected individuals, disseminated MAC historically has occurred almost exclusively in patients with a CD4 count <50 cells/µL.(1,2) Colonization of the respiratory or gastrointestinal (GI) tract by MAC can occur without evident morbidity; however, MAC colonization of these sites indicates that patients are at increased risk for developing disseminated MAC infection.(3) Combination antiretroviral therapy (ART) has been associated with reductions in AIDS-related mortality, days of hospitalization, and the incidence of new OIs.(4) However, there have been numerous reports of aberrant clinical presentations of MAC since the introduction of combination ART.(5-10)

There is convincing evidence that optimal treatment of disseminated MAC infection improves survival and quality of life.(11) Whether MAC can be eradicated once disseminated infection has been established is uncertain, but there have been recent studies that suggest secondary prophylaxis may be discontinued in patients who have had sustained immunologic recovery.(12-14) The treatment regimen for disseminated MAC should include at least 2 drugs to prevent emergence of resistance. The most efficacious drugs for treatment are the newer macrolide antibiotics, clarithromycin and azithromycin.(15-17) The preferred treatment regimen should include 1 of these agents along with ethambutol and/or rifabutin.(18-21) Efficacy results of randomized, controlled trials indicate that clarithromycin, azithromycin, or rifabutin prophylaxis should be administered to prevent MAC disease in all HIV-infected patients with a CD4 count <50 cells/µL.(22-25)

MAC consists of M avium, Mycobacterium intracellulare, and other species of Mycobacterium that have not been classified. In the past, the separation of M avium from M intracellulare was difficult as it relied on serotyping. Therefore, it was common to refer to them as M avium-intracellulare or MAI. Of the 28 known seroagglutination types, types 1 to 6, 8 to 11, and 21 are M avium; types 7, 12 to 20, and 25 are M intracellulare; and types 22 to 24 and 26 to 28 still have not been classified.(26) Using DNA probe techniques, it is now possible to rapidly identify M avium and M intracellulare.(27) In most cases, AIDS-related disseminated MAC infection is caused by M avium species.

MAC strains vary in pathogenicity, which may explain why M avium types 4 and 8 are the most common organisms to infect patients with AIDS.(28,29) Markers of the pathogenicity of the organism include the presence of plasmid, antibiotic susceptibility patterns, restriction fragment length polymorphism patterns, and multilocus enzyme electrophoretic patterns.(30-32)

The source of MAC infection in HIV-infected patients remains uncertain. Water and soil have been implicated as possible environmental sources of MAC; in one well-documented nosocomial outbreak, the hospital water supply was the source of a MAC strain with the same restriction fragment length polymorphism pattern as that infecting a group of hospitalized persons with AIDS.(33)

The risk of developing disseminated MAC infection has inversely correlated with absolute CD4 lymphocyte count values. In one observational study, the 1-year actuarial incidence for developing MAC bacteremia was 3% for patients with a CD4 count in the 100-199 cells/µL range and 39% for patients with a CD4 count <10 cells/µL.(2) The study also revealed that patients with advanced HIV disease had a linear increase in the risk of developing MAC bacteremia over time, and the risk of developing disseminated MAC infection for patients surviving for 30 months after being diagnosed with AIDS was 50%. Other observational studies and prophylaxis trials yielded consistent findings: 30-50% of patients with CD4 counts <50 cells/µL will, in the absence of specific antimicrobial drug prophylaxis, eventually develop disseminated MAC.(1,23,25)

No apparent difference exists in the frequency of disseminated MAC infection according to age, gender, race, risk factor for HIV transmission, initial AIDS-defining condition, or geographic region.(34)

The mode of acquisition of disseminated MAC infection is incompletely understood. Current evidence, however, supports the hypothesis that this disease results from recent environmental acquisition of MAC through either the GI or respiratory tract rather than from reactivation of latent MAC infection.

First, the occurrence of MAC dissemination does not vary among age groups.(34) Thus, if reactivation of latent infection were the mechanism of acquisition, the incidence of dissemination would increase with age, much like in tuberculosis, a disease primarily caused by reactivation of latent infection.

Second, absence of antibody response to MAC in HIV-infected patients with disseminated MAC infection also suggests recent primary infection, because HIV-infected patients lack the ability to mount antibody responses to other primary infections.(35,36) Third, reports of MAC colonization of the GI or respiratory tract preceding the occurrence of disseminated MAC infection suggest that the organism is first acquired from the environment and subsequently disseminated.(3,37)

Finally, the incidence of disseminated MAC among patients with AIDS has substantially exceeded the incidence of MAC infection in the general population, as determined by MAC-specific skin testing.(38)

MAC dissemination requires a susceptible host. By 1984, only 37 cases of disseminated MAC infection in a non-HIV-infected host were reported, mostly in patients with identifiable immunologic defects.(39) In the non-HIV-infected population, MAC infection generally presents as a localized infection: cervical lymphadenitis in children, a slowly progressive fibrocavitary pulmonary disease in middle-aged men with abnormal pulmonary architecture (such as emphysema, bronchiectasis, pulmonary fibrosis, and pneumoconiosis), or right middle lobe or lingula bronchiectasis in women who have a narrow anteroposterior chest diameter, pectus excavatum, scoliosis, or mitral valve prolapse.(40,41)

Surprisingly, histopathologic examination of liver, spleen, bone marrow, or intestine from patients with AIDS-related disseminated MAC shows high-grade infection yet a lack of inflammatory infiltrate or tissue necrosis. In contrast, histopathologic examination of tissue from patients with AIDS who have localized, nondisseminated MAC infection demonstrates marked inflammatory reaction and tissue destruction similar to the histopathologic findings observed in association with non-HIV-associated MAC lung or lymph node infection.

The lack of inflammatory response in patients with AIDS who have disseminated MAC is consistent with the hypothesis that the key immunologic defects associated with dissemination are defective macrophage killing of phagocytized MAC and aberrant cytokine responses, including abnormally low levels of tumor necrosis factor, gamma interferon, and interleukin-12.(42-46)

As a result of this permissive immunologic environment, MAC infection becomes widespread and high grade. In the majority of autopsies performed, MAC has been isolated from spleen, lymph nodes, liver, lung, adrenals, colon, kidney, and bone marrow. The magnitude of mycobacteremia can be as high as 10,000 organisms per milliliter of blood and 10 billion organisms per gram of bone marrow, spleen, lymph node, and liver tissue obtained at autopsy.(47)

As described previously, disseminated MAC infection occurs in patients with AIDS who have CD4 counts <50 cells/µL. Patients with disseminated MAC infection frequently have nonspecific symptoms, signs, and laboratory abnormalities. In many instances, worsening of chronic constitutional symptoms alone may reflect disseminated MAC infection and thus may be attributed incorrectly to progression of other HIV-related illnesses. Patients most commonly report persistent fever, night sweats, fatigue, weight loss, and anorexia. Abdominal pain or chronic diarrhea may result from involvement of retroperitoneal lymph nodes or gut mucosa, respectively. Hepatosplenomegaly, lymphadenopathy, and (rarely) jaundice also may be present. Anemia, which can be severe, is the most common laboratory abnormality, and leukopenia, elevated alkaline phosphatase levels, or low albumin occur in some patients. Intraabdominal adenopathy was demonstrated using abdominal computed tomography (CT) in 14 of 17 patients with disseminated MAC infection.(48)

In a prospective natural history study of MAC bacteremia conducted at San Francisco General Hospital, investigators interviewed patients with CD4 counts <50 cells/µL regarding symptoms and evaluated results of laboratory tests conducted as part of the diagnostic evaluation for disseminated MAC infection.(49) A history of fever for >30 days, a hematocrit <30%, or a serum albumin level <3.0 g/dL were sensitive predictors of MAC bacteremia. Severe fatigue, diarrhea, weight loss, neutropenia, and thrombocytopenia played no role in discriminating between those who were subsequently found to be blood culture positive and those who were negative for MAC.

More unusual manifestations of disseminated MAC infection include palatal and gingival ulceration, septic arthritis and osteomyelitis, endophthalmitis, pericarditis, and massive GI hemorrhage.(50-54)

Mycobacterial culture of peripheral blood is a sensitive and easy method for diagnosing disseminated MAC infection. Mycobacterial blood culture establishes the diagnosis in 86-98% of cases in which disseminated MAC infection is confirmed at autopsy.(55,56) One blood culture identifies 91% of patients with MAC bacteremia, a second blood culture increases the identification rate to 98%.(57) Therefore, obtaining paired or more than 2 sequential blood specimens for culture to diagnose MAC bacteremia is unnecessary.(58)

The preferred culture method includes lysis of peripheral blood leukocytes to release intracellular mycobacteria followed by inoculation onto solid media (eg, Lowenstein-Jensen, Middlebrook 7H11 agar) or into radiometric broth.(59,60) Using the radiometric detection system, mycobacteremia can be detected in as few as 6-12 days, whereas 15-40 days are required with solid media. In addition, DNA probes can identify MAC species within 2 hours once sufficient mycobacterial growth has occurred in radiometric broth or on solid media.(27)

Biopsies from other normally sterile body sites also can prove diagnostic. Stains of biopsy specimens from bone marrow, lymph node, or liver may demonstrate acid-fast organisms or granuloma weeks before positive blood culture results are obtained.(61,62) It is also possible that culture of bone marrow, lymph node, or liver could be more sensitive than blood culture.

MAC is not killed by any standard antituberculous drug except ethambutol at concentrations achievable in plasma. Yet, one half or more of MAC strains can be inhibited by achievable concentrations of rifabutin, rifampin, clofazimine, cycloserine, amikacin, ethionamide, ethambutol, azithromycin, clarithromycin, ciprofloxacin, or sparfloxacin. Unfortunately, drug levels necessary to kill MAC in vitro (minimum bactericidal concentration) are from 8 to >32 times that of inhibitory levels.(63) Whereas combinations of antimycobacterial agents have shown in vitro inhibitory synergism, bactericidal synergism has been more difficult to demonstrate.(63,64) In addition, for in vivo killing, drugs must penetrate macrophages as well as the MAC cell wall. In animal models of disseminated MAC infection, however, both single and combination antimycobacterial regimens have reduced mycobacterial colony counts by several logs and have improved survival.(65,66)

In vivo data on microbiologic efficacy against MAC have been most impressive with the new macrolides, particularly clarithromycin. The focus of one multicenter, randomized, placebo-controlled trial was clarithromycin monotherapy, in dosages of 500, 1,000, and 2,000 mg, all twice daily, in patients with previously untreated disseminated MAC. The investigators reported a median >2 logs decrease in colony-forming units from blood culture specimens, representing a more potent microbiologic effect than had been reported in earlier treatment trials.(15) This microbiologic effect was accompanied by significant clinical improvement in symptoms and quality-of-life indices.

A globally beneficial dose-response effect was not observed, however. Unacceptably high GI toxicity occurred with administration of the 2,000-mg twice daily dosage. Although the 1,000-mg twice-daily dosage had greater microbiologic efficacy than the 500-mg twice-daily dosage, there was actually a trend toward increased mortality in association with the 1,000-mg dosage (13 vs 3 deaths on study in the 1,000- and 500-mg dosing arms, respectively).(15) Not surprisingly, drug resistance emerged after 2 to 3 months of monotherapy in this trial, affecting approximately 50% of patients in all dosing arms. Hence, despite a growing consensus that the currently optimal agent for disseminated MAC therapy is clarithromycin, the macrolide must be combined with other antimycobacterial agents to prevent or at least delay emergence of resistance, which typically is associated with clinical deterioration. A subsequent study confirmed that the 1,000-mg twice-daily dosage of clarithromycin has an adverse effect on survival compared with the 500-mg twice-daily dosage (17/40 deaths [43%] with 1,000 mg twice daily vs 10/45 [22%] with 500 mg twice daily).(67)

Azithromycin is another effective macrolide for MAC treatment. In a trial of azithromycin monotherapy, patients with newly diagnosed positive MAC blood cultures received randomized administration of 1 of 2 doses of azithromycin, 600 or 1,200 mg daily. At 6 weeks, approximately one half of blood cultures were sterile in both groups, with a mean reduction in mycobacteremia of 2.0 and 1.55 logs, respectively (p > .05).(16) Thus, antimycobacterial efficacy was similar to the results with clarithromycin monotherapy. Also, as with clarithromycin, the higher dosage was associated with increased GI intolerance.

Nonmacrolide antimycobacterial agents have been evaluated in several randomized, controlled trials. In one trial, investigators compared the microbiologic efficacy of 4-week monotherapy regimens of rifampin, ethambutol, or clofazimine in patients with previously untreated disseminated MAC.(68) Results showed that only ethambutol effected a statistically significant reduction in blood MAC colony-forming units, suggesting that ethambutol might be the most potent of these 3 antimycobacterial agents.

Data regarding rifabutin treatment for disseminated MAC also have been particularly promising. Interestingly, in the 1980s, when this drug (100-300 mg/day) was administered in a combination regimen with clofazimine as a treatment for MAC, it was found to be ineffective. However, in a 1994 report that described a randomized, placebo-controlled trial in which patients with newly diagnosed disseminated MAC were blindly assigned to receive clofazimine/ethambutol or clofazimine/ethambutol/rifabutin (600 mg/day), approximately half of the patients receiving the rifabutin-containing regimen had a >2 log decrease in blood MAC colony-forming units or sterilization of the blood compared with none of those receiving only clofazimine/ethambutol.(19)

The long-term clinical benefit of combination regimens including both macrolide and nonmacrolide agents for treatment of disseminated MAC was confirmed in a randomized multicenter trial conducted by the Canadian HIV Trials Network. This study included 187 evaluable patients with MAC mycobacteremia who received randomized administration of a regimen of clarithromycin 1,000 mg twice daily, rifabutin 300-600 mg once daily, and ethambutol 15 mg/kg/day or a regimen of ciprofloxacin 750 mg twice daily, rifampin 600 mg once daily, clofazimine 100 mg once daily, and ethambutol 15 mg/kg/day.(11) The in vivo quantitative antimycobacterial effect was significantly better with the macrolide-containing regimen, as was median survival (8.6 vs 5.2 months; p < .001).

Although many studies have yielded positive results, the composition of the optimal treatment regimen for disseminated MAC is still unclear. Certainly, 1 drug should be a macrolide, such as clarithromycin or azithromycin. No significant difference in mortality was noted in a randomized, double-blind, multicenter, international trial comparing azithromycin 600 mg orally once a day with clarithromycin 500 mg orally twice a day in an ethambutol-based regimen.(17) Evidence from a 1995 report describing a trial of several clarithromycin-containing combination regimens suggests that clofazimine compares poorly with ethambutol as a second drug added to clarithromycin.(18) Moreover, a randomized clinical trial that added clofazimine to clarithromycin plus rifabutin found higher mortality in the clofazimine-containing arm (62% with clofazimine vs 38% without clofazimine; p = .012).(69) Also, a randomized trial examining a 4-drug, oral antimycobacterial regimen with or without parenteral amikacin showed no clinical benefit associated with amikacin therapy.(70) An open-label, prospective, randomized trial comparing the efficacy and safety of clarithromycin plus ethambutol, rifabutin, or both for the treatment of disseminated MAC was conducted by the AIDS Clinical Trials Group (ACTG). A total of 203 patients were enrolled; 53 received clarithromycin and ethambutol, 50 received clarithromycin and rifabutin, and 57 received all 3 drugs. A complete microbiologic response, as defined by sterile blood cultures at 12 weeks, was seen in 40%, 42%, and 51%, respectively. A significant improvement in survival was observed in the 3-drug arm.(20) However, another randomized, placebo-controlled study evaluating the addition of rifabutin to a regimen of clarithromycin and ethambutol noted no difference in the microbiologic response or survival benefit between the 2 arms.(21) The latter study did detect a significant reduction in the development of clarithromycin resistance in the arm containing all 3 drugs (2% vs 14%), suggesting a possible benefit. Table 1 summarizes the best treatment options: clarithromycin and ethambutol, with or without rifabutin.

It appears safe to discontinue antimycobacterial therapy in patients with AIDS who have been treated for at least 12 months with a macrolide-containing regimen and have a sustained CD4 count >100 cells/µL for at least 3 months as a result of combination ART. (See detailed discussion in the following section on prophylaxis.)

Because 40% of patients with advanced HIV disease are likely to develop disseminated MAC, it makes sense to develop a strategy for preventing this disease in patients at risk. Few specific risk factors, other than a low CD4 lymphocyte count, have been defined. Therefore, any current prophylactic strategy must be applied to the entire population at risk, specifically, those patients with <50 CD4 cells/µL.

Results of combined analysis of 2 randomized, placebo-controlled trials of rifabutin prophylaxis, which included more than 1,000 patients with advanced HIV disease, showed that 300 mg/day of rifabutin reduced the incidence of mycobacteremia by one half.(25) Of interest, patients who received rifabutin and subsequently developed mycobacteremia had blood MAC isolates that retained susceptibility to rifabutin. Neither trial alone, nor the combined analysis, however, demonstrated that rifabutin significantly reduced mortality. Combined analysis of both trials revealed an increased incidence of fever, fatigue, anemia, elevated alkaline phosphatase levels, and hospitalizations in patients who received placebo compared with those who received rifabutin. Patients assigned to placebo in these trials, however, tended to have lower absolute CD4 counts, which may have accounted for some of the greater morbidity.

A placebo-controlled study provided the first evidence that MAC prophylaxis could improve survival of patients with late-stage HIV disease. A total of 682 patients with advanced HIV disease were randomly assigned to receive either clarithromycin (500 mg) or placebo twice daily. During a median 10-month follow-up period, only 6% of clarithromycin-assigned patients developed mycobacteremia compared with 16% of placebo-assigned patients (p < .001).(23) More importantly, median survival was significantly longer for clarithromycin-assigned patients than for placebo-assigned patients (8.6 vs 5.2 months; p = .001). Dose-limiting drug toxicity occurred in 8% of the clarithromycin-assigned patients and 6% of the placebo-assigned patients, respectively.

A second trial provides even more convincing evidence of the long-term benefit of clarithromycin prophylaxis for MAC in patients with very advanced HIV disease. This trial was conducted by the National Institute of Allergy and Infectious Diseases ACTG and Community Program on Clinical Research on AIDS (ACTG 196/CPCRA 009). A total of 1,216 patients with a median absolute CD4 count of 28 cells/µL were randomly assigned to receive rifabutin (300-450 mg/day), clarithromycin (500 mg twice daily), or a combination of these agents.(22) Median follow-up of patients enrolled in ACTG 196/CPCRA 009 was 589 days (more than 19 months), longer than in any other MAC prophylaxis trial ever performed; overall survival was approximately 50%. At the conclusion of the trial, deaths in the respective groups were: clarithromycin, 42%; rifabutin, 43%; and the combination, 46%.

In an intent-to-treat analysis, only 9% of patients assigned to clarithromycin developed disseminated MAC, compared with 15% assigned to rifabutin (p < .001). The combination of clarithromycin and rifabutin was not significantly better than clarithromycin alone. Data from previous natural history studies and placebo-controlled prophylaxis trials would predict, in this trial group, a 36% incidence of disseminated MAC at 19 months in the absence of any MAC prophylaxis. Thus, long-term clarithromycin prophylaxis administered to 100 patients would be expected to prevent 27 cases of disseminated MAC over the median lifetime of the population at risk (36 expected disseminated MAC cases minus 9 observed cases per 100 patients, assuming a median survival of 19 months).

In a third macrolide prophylaxis trial, conducted by the California Collaborative Treatment Group (CCTG), 693 patients with <100 CD4 cells/µL were randomly assigned a regimen of azithromycin (1,200 mg once per week), rifabutin (300 mg/day), or a combination of these drugs.(24) Median follow-up period was 514 days. The incidence of disseminated MAC was significantly lower with azithromycin (13.9%) or the combination (8.3%) when compared with rifabutin (23.3%). Dose-limiting drug toxicity occurred in 13%, 23%, and 16% of the 3 groups, respectively.

Fortunately, breakthrough with macrolide-resistant mycobacteremia has been a rare event in the prophylaxis trials just described; only 2-3% of patients given clarithromycin and only 1% of patients given azithromycin had "breakthrough" mycobacteremia with MAC isolates that were resistant to these drugs. In comparing the clarithromycin and azithromycin trial results, note that the CCTG trial patients began with a higher median CD4 count and had a shorter follow-up period; therefore, they were probably at lower risk. Nevertheless, the rate of breakthrough disseminated MAC was as low on a regimen of clarithromycin alone (9%) in the ACTG 196/CPCRA 009 trial as it was on a regimen of the azithromycin/rifabutin combination (8.3%) in the CCTG trial.

Two randomized, controlled studies have compared azithromycin with placebo in patients with CD4 counts increased to >100 cells/µL on ART and no prior history of disseminated MAC.(71,72) Specifically, in a study conducted by the CPCRA, a total of 520 patients were enrolled with a median CD4 count of 230 cells/µL at entry. There were no cases of confirmed MAC in either the azithromycin arm or the placebo arm over a 12-month period.(71) In ACTG 362, a study enrolling 643 patients, there were 2 cases of MAC observed among the 321 patients assigned to placebo and no cases of MAC occurring in those patients who remained on azithromycin, yielding an incidence of 0.5 events per 100 person-years.(72) In both of these studies, discontinuation of azithromycin prophylaxis was not found to increase the incidence of MAC disease significantly during a median of 12-16 months of follow-up, suggesting that it is safe to discontinue primary prophylaxis in those patients who have had a CD4 T-cell count recovery attributed to ART.

In the initial pilot study evaluating the safety of discontinuing secondary prophylaxis for disseminated MAC, investigators at San Francisco General Hospital prospectively followed 4 patients with a known history of the disease. Antimycobacterial therapy was discontinued without evidence of recurrence after 8-13 months of follow-up.(12) In the larger, prospective ACTG 393 trial, a total of 48 patients discontinued antimycobacterial therapy.(13) All patients had been on a macrolide-based regimen for at least 12 months, and on ART with a CD4 count >100 cells/µL for at least 16 weeks. Of the 48 subjects, 47 remained MAC free for a median of 77 weeks off therapy. One patient developed a localized rib osteomyelitis due to MAC 16 months after stopping secondary prophylaxis. In a French retrospective study, 3 of 26 patients experienced a MAC relapse.(14) One patient was extremely immunosuppressed with a CD4 count of <50 cells/µL at the time of relapse, and the other 2 developed atypical bone infections with CD4 counts of 126 and 160 cells/µL, respectively. In ACTG 362, a primary prophylaxis withdrawal study,(72) there were 2 cases of MAC osteomyelitis reported in the placebo arm. Although the rarity of these events suggests that it is safe to discontinue primary and secondary prophylaxis, clinicians do need to be aware that atypical manifestations of MAC may occur in patients having sustained elevations of CD4 T cells.(73) In patients with increased CD4 counts on ART who develop atypical manifestations of MAC, it may be prudent to treat for a prolonged period of 12-18 months and then continue chronic suppressive therapy for life. It is also critical to remember to reinitiate MAC prophylaxis if immunologic and/or virologic failure results in a decrease in the CD4 T-cell count. At this time, the absolute CD4 T-cell count indicating reinitiation is unknown, although the U.S. Public Health Service guidelines still recommend a CD4 count of <50 cells/µL to start prophylaxis.

There is no standard definition for immune reconstitution inflammatory syndrome (IRIS) nor are there definitive management strategies.(74-76) IRIS tends to occur more frequently in individuals receiving ART for the first time who experience a rapid decline in HIV viral load.(74) However, although there have been attempts to define IRIS based upon CD4 T-cell increases and viral load reductions, development or worsening of symptoms soon after the initiation of HIV treatment (so-called early or paradoxical reactions) may occur prior to the establishment of laboratory-based evidence of "immune reconstitution". Perhaps the best-described cases of IRIS have consisted of atypical manifestations thought secondary to subclinical MAC disease in the setting of ART.(5) Subsequent studies described late manifestations occurring in individuals who discontinued either primary or secondary prophylaxis.(13,14,73) IRIS has been reported more frequently among those who begin ART in the setting of an acute opportunistic infection (hence the need to conduct studies that evaluate the optimal time to start ART). Other uncertainties include whether IRIS results from actual growth of the organism and whether pathogen-specific therapy is indicated. In some settings, antiinflammatory agents may be preferred. It is recommended that patients presenting with presumed IRIS be treated on an individual basis with consideration of the use of nonsteroidal antiinflammatory drugs (NSAIDs) or steroids.

For more information on IRIS, see the HIV InSite Knowledge Base chapter Clinical Implications of Immune Reconstitution in AIDS.

The advent of antimycobacterial macrolide therapy has greatly improved both prophylaxis and treatment of disseminated MAC infection. Although questions remain about which particular regimens are optimal, several general recommendations can be made:

Suspect disseminated MAC in patients with CD4 <50 cells/µL and nonspecific symptoms, signs, and laboratory abnormalities.

Mycobacterial culture of peripheral blood is a sensitive method for diagnosing disseminated MAC. A single blood culture identifies 91% and 2 cultures identify 98% of cases of MAC bacteremia. Stains and culture biopsies from other normally sterile body sites (eg, bone marrow, lymph node, or liver) also may be useful.

Patients with CD4 counts <50 cells/µL who exhibit no clinical evidence of active mycobacterial disease should receive prophylaxis with either clarithromycin (500 mg twice daily) or azithromycin (1,200 mg weekly); the latter could be coadministered with 300 mg rifabutin daily. MAC prophylaxis should be discontinued in adult patients without a history of MAC disease whose CD4 counts remain >100 cells/µL for at least 3 months on ART.

Optimal treatment should begin with clarithromycin (500 mg twice daily) plus ethambutol (approximately 15 mg/kg/day). Rifabutin may be added to this regimen, but the exact rifabutin dosing depends on other concomitant medications that might result in drug interactions. Three-drug anti-MAC therapy is preferred for individuals who do not plan to start or cannot be prescribed an effective antiretroviral regimen.

It appears safe to withdraw primary and secondary prophylaxis in patients who have sustained CD4 counts >100 cells/µL for at least 3-6 months on ART. Prophylaxis should be reinitiated if the CD4 T-cell count declines.