Skip Navigation

Correlation of Quantitative Bone Marrow and Blood Cultures in AIDS Patients with Disseminated Mycobacterium avium Complex Infection

  1. Richard Hafner1,
  2. Clark B. Inderlied4,
  3. Dolores M. Peterson5,
  4. David J. Wright2,
  5. Harold C. Standiford3,
  6. George Drusano6 and
  7. Katherine Muth2 The Division of AIDS Treatment Research Initiative Protocol 007 Study Groupa
  1. 1Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda
  2. 2Westat, Rockville
  3. 3Institute of Human Virology, University of Maryland School of Medicine, and Department of Veterans Affairs Medical Center, Baltimore, Maryland
  4. 4Children's Hospital of Los Angeles and Department of Pathology and Laboratory Medicine, University of Southern California School of Medicine, Los Angeles
  5. 5University of Texas Southwestern Medical Center, Dallas
  6. 6Albany Medical College, Albany, New York
  1. Reprints or correspondence: Dr. Richard Hafner, Division of AIDS, National Institute of Allergy and Infectious Diseases, 6700-B Rockledge Dr. MSC 7624, Bethesda, MD 20892-7624 (rhafner{at}niaid.nih.gov).
 
Next Section

Abstract

The relationship between Mycobacterium avium complex (MAC) infection of blood and bone marrow was studied in human immunodeficiency virus—infected patients before and during treatment. Quantitative cultures were obtained at baseline from 17 persons with newly detected MAC bacteremia. Serial blood cultures were obtained, and a second bone marrow sample was obtained at 4 or 8 weeks. At baseline, the median MAC load in bone marrow core samples was 3 log10 higher than in blood. Bone marrow MAC loads ranged widely (866–847,315 cfu/g), and no significant correlation was found between MAC load in blood and that in bone marrow core samples. MAC loads in bilateral bone marrow biopsy samples from 7 subjects were highly correlated. MAC loads declined in blood and bone marrow at similar rates during therapy, but blood was sterilized before bone marrow. Length of survival was inversely associated with initial bone marrow core MAC load but not with blood MAC load. Initiation of treatment when tissue MAC load is low may increase the likelihood of favorable clinical outcome.

Disseminated Mycobacterium avium complex (MAC) is a debilitating human immunodeficiency virus (HIV)—related opportunistic infection [1–4]. Knowledge of the natural history of disseminated MAC (DMAC) infection is incomplete, and the relationship between bacteremia and tissue infection is not clear. Autopsy studies done >10 years ago [5] suggested that the mycobacteremia associated with DMAC disease is the result of “spillover” from heavily infected tissues. This theory was based on estimates of MAC tissue burden as high as 1010 cfu/g [6], whereas concomitant mycobacteremia levels in the same patients typically were 1–104 cfu/mL. However, later studies suggested that some patients have transient bacteremia before or during tissue infection [7] and that mycobacteremia precedes widespread infection of tissues [8].

Clinical trials of multidrug regimens [9–11] and more recent trials that included macrolides [12–15] showed that MAC treatment effectively reduces the number of organisms in the blood and improves MAC-related symptoms. More recently, macrolide-based combinations have improved survival compared with older regimens [16]. However, the long-term outcomes remained poor, and relapse still was common [17–20]. MAC infection remaining in tissue after the sterilization of blood is the probable source of relapses and may adversely influence long-term clinical responses. Blood culture results have not been clinically correlated with MAC loads in tissues before or during treatment, although an autopsy study suggested good correlation between levels of bacteremia and tissue infection after several months of therapy [21].

The purpose of this study was to determine the relationship between the organism loads of MAC in blood and bone marrow at the time bacteremia is detected and during therapy. In addition, a substudy evaluated the uniformity of infection in bone marrow by comparison of quantitative culture results from samples obtained at 2 biopsy sites. Bone marrow biopsy was utilized in this study because it provides a relatively safe method to obtain a tissue sample with a high concentration of reticuloendothelial cells for quantitative culture.

Previous SectionNext Section

Patients and Methods

Patient population

Eligible subjects were HIV-infected adults (⩽ 18 years old) in whom MAC bacteremia was first identified from samples obtained within 12 weeks before study entry. The following parameters were required within 7 days before study entry: serum creatinine <3.0 mg/dL, bilirubin <2.0 mg/dL, aspartate aminotransferase (AST)/alanine aminotransferase <5 times the upper limit of normal, absolute neutrophil count >750 cells/mm3, platelet count >50,000 cells/mm3, and hemoglobin >7.0 g/dL. Patients could not have received any drug with activity against MAC for >10 days during the 8 weeks before study entry or any treatment or prophylaxis for MAC between the collection of the first MAC-positive blood sample and study entry. Persons who had received immunomodulators or cytotoxic chemotherapy within 8 weeks of study entry were excluded. Acute therapy for AIDS-related opportunistic infections or any other acute medical illnesses must have been completed 4 weeks before enrollment. Women were excluded if they were pregnant or breast-feeding.

Study methods

At baseline, 2 bone marrow core samples, 3–5 mL of bone marrow aspirate (from the same entry site), and 2 blood cultures were obtained from each patient. Patients were randomly assigned to undergo a second bone marrow biopsy after 4 or 8 weeks of treatment. At day 1 and weeks 1, 2, 4, 6, 8, 12, and 24, quantitative blood cultures and susceptibility testing were repeated, and plasma was obtained and stored at −70°C. At these visits and at weeks 36 and 48, patients completed a history and had a physical examination with routine hematology and chemistry tests. Patients kept a daily diary to record evening temperature, presence of night sweats, and antipyretic use. They were also asked to participate in a substudy to assess variation in quantitative MAC culture results between bone marrow biopsy samples obtained from the left and right iliac crests at baseline. HIV-1 load was determined from frozen plasma samples after the study ended.

Study therapy

After the baseline procedures, subjects received 500 mg of clarithromycin twice daily (provided by Abbott Laboratories, Abbott Park, IL) and 15 mg/kg of ethambutol daily, to a maximum of 1200 mg/day (provided by Lederle Laboratories, Pearl River, NY), for 48 weeks. At the discretion of the treating physician, rifabutin (300 mg/day) could be added to the treatment regimens of patients whose baseline MAC-associated signs and symptoms did not improve in the first 8 weeks.

Medication compliance

Compliance was assessed by counting medications returned at each study visit and was confirmed by medical event monitoring system electronic medication compliance container caps. For the purposes of analysis, patients were considered compliant if they received >60% of each study drug.

Cultures

All specimens for culture were shipped overnight to a central reference laboratory. Ten milliliters of blood was collected for culture in a sodium polyanethol sulfonate tube. Three to 5 mL of blood was inoculated into a blood culture vial (Bactec 13A; Becton Dickinson, Sparks, MD). The remaining blood was treated with sodium deoxycholate to lyse blood cells. The lysate was centrifuged at 3500 g and resuspended in Middlebrook 7H9 broth (Difco, Detroit) with serum albumin. Serial dilutions were made in 7H9 broth, and aliquots were inoculated in duplicate onto Middlebrook 7H11 agar supplemented with oleic acid—albumin-dextrose and catalase (Difco). The colonies on the plate were counted, and viable bacteria per milliliter of the original blood sample were calculated as colony-forming units per milliliter. The Bactec 13A media were incubated at 35°C and monitored for growth for 6 weeks with a Bactec 460 instrument (Becton Dickinson). Cultures were positive when the Bactec growth index was 30 and a smear of the growth medium was positive for acid-fast bacilli (AFB). A subculture was made onto 7H11 agar for species identification and susceptibility testing and was incubated at 35°C in 5% CO2 for 8 weeks.

A portion of the bone marrow aspirate (3–5 mL) was placed in a 1.5-mL pediatric isolator tube (Wampole, Cranbury, NJ) for quantitative culture, and the rest was placed in 2 mL of 10% buffered formalin for histology. One bone marrow core sample was weighed and placed in nonbacteriostatic saline, and the second core was placed in 2 mL of fresh 10% buffered formalin. The core sample in saline was macerated, diluted in 7H9 broth, and inoculated in duplicate onto 7H11 agar. In addition, 0.1 mL of the suspension was inoculated into Bactec 12B media. The bone marrow cultures were incubated and examined as previously described [22]. Mycobacteria were identified with DNA probes (GenProbe, San Diego) or by conventional methods [23].

Susceptibility testing

MAC isolates were tested against clarithromycin (2–32 µg/mL) and ethambutol (1–16 µg/mL) by use of a broth radiometric (Bactec) microdilution assay at pH 6.9 [24]. Each drug was tested at 5 concentrations, and the MIC was defined as the lowest concentration of drug that inhibited 99% of growth when a starting inoculum of ∼5 × 104 cfu/mL was used; MAC 101 type 1 was used for quality control. The breakpoint for clarithromycin resistance was defined as 32 µg/mL [24].

MAC strain identification

Strains of MAC were identified by restriction fragment length polymorphism analysis and field inversion electrophoresis, by use of a procedure described by Arbeit et al. [25] and modified by Nash and Inderlied [26].

Histopathology studies

Bone marrow core samples in formalin were decalcified with formic acid, paraffin-embedded, and processed. Sections (4 µm) were stained with Mayer's hematoxylin-eosin and Kinyoun AFB stains. In addition, an immunoperoxidase (IP) stain for mycobacteria with rabbit polyclonal antibodies to M. paratuberculosis (a subspecies of M. avium; Dako, Carpinteria, CA) was used by standardized technique [27]. All sections were examined by a hematopathologist to determine specimen quality, cell types, presence of granulomas, and any other abnormalities. The presence of AFB- or IP-staining material was graded 1 + to 4 +.

HIV load

Plasma HIV-1 RNA levels were measured in a central laboratory (Laboratory Corp. of America, Research Triangle Park, NC), by use of a reverse-transcriptase polymerase chain reaction kit (Amplicor, HIV-1 MONITOR assay; Roche Molecular Systems, Branchburg, NJ) with a lower limit of quantitation of 400 copies/mL.

Statistical methods

The associations between categorical variables and between continuous variables were assessed by use of Fisher's exact χ2 test and Spearman's correlation coefficient, respectively. The 7H11 plate cultures negative for MAC were scored as 0.1 cfu/mL if Bactec-positive and as 0.01 cfu/mL if Bactec-negative. Rates of decline were estimated from a repeated-measures model that included a random intercept effect for each subject [28]. Median time to events was estimated by the Kaplan-Meier method. The log rank test was used to assess potential covariates [29]. All computations were done by use of SAS software [30].

Previous SectionNext Section

Results

Baseline Results

Patient characteristics

Eighteen patients were enrolled between July 1993 and March 1995 at 4 participating centers. One person was excluded as ineligible because the baseline blood culture was negative. A summary of patient characteristics is shown in table 1. Eleven patients were enrolled on the basis of routine surveillance blood cultures, and, in general, symptoms of DMAC disease were mild or absent at baseline. Fever was the most common symptom reported, but only 4 of 17 subjects had a measured temperature ⩽37.8°C (range, 37.8–38.7) at study entry. At baseline, the mean CD4+ lymphocyte count was 15 cells/mm3, and the mean HIV-1 RNA concentration was 5.08 log10 copies/mL. At study entry, 9 of 17 patients were receiving a single nucleoside reverse-transcriptase inhibitor; the remaining subjects were not receiving antiretroviral therapy.

Figure 1
View larger version:
Figure 1

Correlation of Mycobacterium avium complex (MAC) load in blood with MAC load in bone marrow core and aspirate specimens at baseline. Quantitative culture results for blood and bone marrow aspirate were significantly correlated (ρ = .53, P = .03), but the results for blood and bone marrow core biopsy specimens were not (ρ = .19, P = .48). MAC loads were higher in both bone marrow core and aspirate specimens than in the corresponding blood specimen for all patients.

Figure 2
View larger version:
Figure 2

Rates of decline of Mycobacterium avium complex (MAC) load in blood, bone marrow core, and bone marrow aspirate specimens during the first 8 weeks of treatment with clarithromycin and ethambutol. Changes in MAC loads are depicted in A–C for each of 11 subjects included in MAC load rate-of-decline analysis. Blood specimens (A) were collected at baseline and every 2 weeks, and bone marrow core (B) and aspirate (C) specimens were obtained at baseline and at either 4 or 8 weeks. D, Rates of decline of MAC loads in blood, bone marrow core, and bone marrow aspirate specimens, as estimated by use of repeated-measures models.

Table 1
View larger version:
Table 1

Demographic and general laboratory data for 17 study patients at baseline.

Microbiology data

MAC was isolated from all baseline blood and bone marrow specimens obtained from the 17 eligible subjects. All isolates were M. avium except for the isolates from 1 patient. M. intracellulare was identified from this patient's samples, and M. avium was also isolated from the day 1 blood culture. The MAC load in 1 patient's baseline bone marrow core sample could not be calculated because of a weighing error that could not be corrected. This patient's bone marrow aspirate MAC load was the lowest of all baseline aspirate cultures. Table 2 shows each patient's quantitative MAC culture results at baseline and at weeks 4 and 8, HIV-1 results at baseline, and clinical status at the end of the study.

Table 2
View larger version:
Table 2

Baseline and follow-up microbiologic and clinical data for individual patients.

The median log10 MAC loads (and ranges) at baseline were 1.57 (−1.00 to 3.14) cfu/mL for blood, 2.68 (0.90–4.20) cfu/mL for bone marrow aspirate, and 4.60 (2.94–5.93) cfu/g for bone marrow core specimens. The median difference between log10 results for baseline bone marrow core and blood cultures was 2.70 (range, 1.46–4.91) cfu/mL or cfu/g. At baseline, the quantitative culture results for bone marrow core and blood samples were not significantly correlated (Spearman's correlation coefficient ρ = .19, P = .48), but the bone marrow aspirate and blood culture results were significantly correlated (ρ = .53, P = .03; figure 1). A significant correlation was found between the bone marrow core and aspirate culture results (ρ = 0.58, P = .02). No statistically significant correlations were found between the core MAC load and biopsy specimen weight or any of the following baseline characteristics: presence of fever or fatigue, weight, Karnofsky score, AST, alkaline phosphatase, CD4 cell count, or HIV-1 RNA concentration. However, the detection of a weak correlation may have been limited because of the small sample sizes.

Bilateral bone marrow biopsies

MAC loads in core biopsy specimens and aspirates obtained at baseline from the right and left iliac crests of 7 patients were compared to determine the uniformity of MAC infection. The same-patient log10 SDs for core and aspirate cultures were +0.31 cfu/g and +0.28 cfu/mL, respectively, corresponding to an approximately 4-fold variation between bilateral samples.

Bone marrow histology

At least 1 core specimen from each patient and 23 (96%) of the 24 total baseline core specimens contained bone marrow elements. No bone marrow elements were present in the baseline aspirates from 9 (53%) of 17 patients and in 13 (54%) of 24 aspirates. Mature, fully formed granulomas were observed in the bone marrow core samples from 2 (12%) of 17 patients. No differences in histopathology were observed between the left and right biopsy specimens. AFB- or IP-staining bacilli were observed in ⩽1 sample from 9 (53%) of 17 patients and in 13 (54%) of the 24 baseline core biopsy specimens. No difference in sensitivity existed between the AFB and IP stains. Both were significantly more likely to be positive when the core MAC load was >15,000 cfu/g (8/10 patients) and negative when the core MAC load was <15,000 cfu/g (5/6 patients; P = .035).

Response to Treatment

Evaluable patients

Fourteen patients were compliant and 3 noncompliant with study therapy, according to study criteria (see Patients and Methods); noncompliant patients were excluded from all response-to-treatment analyses. Among the 14 compliant subjects, compliance was 62%–100% (median, 93%) for clarithromycin and 74%–100% (median, 100%) for ethambutol. Among the 3 noncompliant patients, compliance for clarithromycin ranged from 12% to 51%. Only 1 subject received a 3-drug regimen (rifabutin was started 4 months into MAC study treatment). Three compliant patients were excluded from the analysis of the rate of decline in MAC load for the following reasons: the MAC load in the baseline core sample of 1 patient could not be calculated because of a weighing error, 1 patient withdrew from the study before the second biopsy, and the MAC load in the second core sample from 1 patient could not be determined because of staphylococcal overgrowth. The baseline characteristics of the 6 excluded patients were not statistically different from those of the 11 patients included in the rate-of-decline analysis.

Rates of decline in MAC loads

Figure 2A–2C depicts the changes in MAC load for each type of specimen obtained from the 11 evaluable subjects. The baseline MAC loads for each type of sample were not significantly different between the randomization groups (table 3). The mean MAC load was greater in blood obtained from the week 4 patients (1.70 log cfu/mL) than in blood from the week 8 patients (0.93 log10 cfu/mL), but this difference was not statistically significant (P = .16). The rates of decline in blood and bone marrow samples were not statistically different between the 2 groups, suggesting that the rates of decline during the first and second 4 weeks of treatment were similar. By use of repeated measures models, including all évaluable patients, the log10 rate of decline and SE over 8 weeks was 0.36 ± 0.03 cfu/mL/week for blood, 0.30 ± 0.07 cfu/mL/week for bone marrow aspirate, and 0.29 ± 0.06 cfu/g/week for bone marrow core (figure 2D). The rates of decline for the 3 specimen types were not statistically different (P = .34) from a common rate of decline (log10 0.34 cfu/g or mL/week) and were independent of the baseline MAC load.

Table 3
View larger version:
Table 3

Mean Mycobacterium avium complex loads (log10 cfu/mL or g) in blood, bone marrow core, and bone marrow aspirate specimens and rates of decline for each randomization group.

Sterilization of blood and bone marrow

Among the 7 évaluable patients in the week 4 bone marrow biopsy group, 2 had sterile blood cultures at week 4, and 6 had sterile blood cultures at week 8, but all 7 remained bone marrow culture—positive at week 4. Although 2 patients in the week 8 biopsy group were not included in the rate-of-decline analysis because MAC load in a bone marrow core sample could not be calculated, these subjects were included in the assessment of sterilization. Among the 6 compliant patients in the week 8 biopsy group, all had sterile blood cultures, but only 2 had sterile bone marrow cultures at week 8. Among all compliant patients, the mean time to sterilization of blood was 7.0 (±0.7 SE) weeks.

Characteristics of MAC isolates

Selected blood and bone marrow isolates were obtained from all subjects at baseline and during therapy, and all isolates from the same person had identical macrorestriction patterns of genomic DNA. The MICs for clarithromycin and ethambutol were determined for>150 isolates. No significant differences in MICs were found between blood and bone marrow isolates collected from the same person during the study. Resistance to clarithromycin (MIC = 32 µg/mL) was not detected, except in a bone marrow isolate obtained from a patient who relapsed at week 40.

MAC loads and levels of HIV-1 RNA

The mean baseline HIV-1 RNA level was 5.08 log10 copies/mL among the 11 persons in the MAC load rate-of-decline analysis. Seven of these subjects were receiving antiretroviral therapy with a single nucleoside reverse-transcriptase inhibitor before enrollment. Two additional patients began antiretroviral monotherapy after study entry, 1 at week 12 and 1 at week 36. The estimated mean log10 rate of change in HIV-1 RNA levels by repeated-measures modeling was 0.01 (±0.02 SE; 95% confidence interval [CI], −0.03 to 0.05) log10 copies/mL/week over the first 8 weeks. No statistically significant correlation was found between individual patient changes in HIV load and individual patient changes in MAC load of bone marrow core, bone marrow aspirate, or blood.

Patient survival

Among the 14 treatment-compliant patients, the mean survival time during the 48-week study period was 37.3 (±2.8 SE) weeks, as estimated by the Kaplan-Meier method. Survival time had a statistically significant inverse association with the baseline MAC load in bone marrow core (P = .02). None of the 7 persons with baseline core MAC loads <20,000 cfu/g (including the 1 patient whose baseline core MAC load was not quantitated but was presumed to be <20,000 cfu/g based on the aspirate results) died during the 48-week study period, although 1 was lost to follow-up after week 38. One patient died 8 months after study completion of central nervous system lymphoma. No evidence of MAC infection was found on autopsy, which included extensive tissue examinations. Of the 7 subjects with baseline core MAC loads >20,000 cfu/g, only 1 was known to have survived beyond week 48. Four died before week 48. Two were lost to follow-up after 3 and 24 weeks, respectively, and both were in severely deteriorating health at their last study evaluation. No statistically significant associations were found between survival time and baseline aspirate or blood MAC loads or baseline HIV-1 RNA levels.

Previous SectionNext Section

Discussion

This study was designed to clinically evaluate the course of DMAC disease before and during antimicrobial therapy by determination of the quantitative relationship between MAC loads in blood and tissue. All subjects had higher MAC loads in bone marrow than in blood, and the MAC load in bone marrow core samples was, on average, 1000-fold higher than that in blood. The presence of higher MAC loads in bone marrow is not surprising, because bone marrow is expected to have a higher concentration of infectable cells. More notably, the study patients had a wide range of bone marrow core MAC loads (866–847,315 cfu/g) at the time of detectable bacteremia. Although some subjects had high MAC loads in bone marrow at the time low-grade bacteremia was detected, several others had relatively low bone marrow loads at the time of sustained bacteremia. These results indicate that some persons do not have overwhelming levels of infection in at least one type of tissue after detection of sustained bacteremia. This finding is in contrast to conclusions based on autopsy studies performed in the 1980s and the associated “spillover theory.”

Also notable is the lack of correlation between MAC loads in bone marrow core and blood samples at baseline. These results indicate that quantitative blood cultures cannot be used to distinguish between patients with high and low tissue MAC loads before treatment. MAC loads in the baseline bone marrow core and aspirate samples were significantly correlated; thus, aspirates could be used as a measure of tissue MAC load if core samples are not available. However, the median aspirate MAC load was >200-fold lower than the core load, and aspirates are not as sensitive as core samples for the detection of very low MAC levels in bone marrow. The higher MAC loads in bone marrow core samples probably reflect a higher concentration of cells likely to be infected with MAC. Ninety-six percent of baseline core samples had identifiable marrow elements, compared with 54% of aspirate samples. Aspirate MAC loads also were significantly correlated with blood MAC loads, probably because aspirates contain a mixture of bone marrow elements and peripheral blood (which was apparent on visual and histologic examination). In contrast, microscopic examination of bone marrow core samples rarely revealed contamination with peripheral blood cells.

Bilateral biopsies were performed at baseline to assess the uniformity of MAC infection in bone marrow. If MAC infection is not uniform, sampling differences could lead to false conclusions concerning the intensity of bone marrow infection and the quantitative relationship between blood and bone marrow MAC loads. Bilateral biopsy specimens were obtained from 7 patients, and the same-patient variation between right and left samples averaged 0.3 log10 cfu/g (∼4-fold). Havlir et al. [31] showed that the results of paired quantitative MAC blood cultures done within a single laboratory were highly reproducible, with an average within-patient variation of ∼1.8-fold. Although the 0.3 log10 cfu/g variation between bilateral core MAC loads is somewhat greater than the expected experimental error, variation of this magnitude should allow clinically relevant use of a single biopsy to assess the intensity of MAC load in bone marrow. However, MAC load in bone marrow may not reflect the intensity of infection in other tissues.

The average rates of decline in MAC load during the first 8 weeks of treatment appear to be similar for bone marrow core and blood cultures (0.29 and 0.36 log10 cfu/g or cfu/mL per week, respectively). The rates of decline in bone marrow and blood were independent of baseline MAC loads, indicating that the population density of bacilli in tissue is not likely to have a large influence on the effectiveness of MAC chemotherapy. The time to sterilization of bone marrow by chemotherapy appears to correlate with the baseline bone marrow MAC load but could not be predicted on the basis of blood culture response. At week 8, only 2/6 patients (both with very low bone marrow core MAC loads) had sterile bone marrow cultures, but all 6 évaluable patients had sterile blood cultures. Although the sterilization of blood occurred relatively rapidly (median, 7 weeks), the sterilization of bone marrow usually occurs considerably later, particularly in persons with higher MAC loads in bone marrow. Patients who have poor or intermittent compliance with therapy before sterilization of tissues are at high risk for relapse and are likely to relapse with a macrolide-resistant MAC isolate.

Histopathologic findings were rare, and well-formed granulomas were uncommon in bone marrow core specimens, regardless of the level of MAC infection. These results are consistent with earlier studies of MAC-infected tissues from HIV-seropositive subjects [32, 33]. In a recent study of 82 bone marrow biopsy specimens from 76 persons, Kilby et al. [34] observed granulomas in only 27% of the culture-positive specimens. In the present study, AFB and IP staining had similar sensitivity for the detection of bacilli in bone marrow core samples. Relatively low levels of marrow infection (<15,000 cfu/g) were not reliably detected by either technique. This finding is consistent with the results of Kilby et al. [34], who found that only 29% of mycobacterial culture—positive bone marrow samples were AFB-positive.

Torriani et al. [8] examined the relationship between blood and tissue MAC infection in an autopsy study of 44 subjects with MAC bacteremia. No evidence of MAC infection was detected by histologic examination of any tissues obtained from 30% of the subjects, and the risk of detectable histologie involvement was related to the duration of bacteremia. The investigators concluded that tissue burden at the time of initial bacteremia may be limited in many persons and that detectable MAC bacteremia may precede widespread tissue disease. Although this autopsy study assessed persons at a much later phase of MAC disease, the results of the present study are in agreement with the first of these conclusions. The two studies reach different conclusions concerning the frequency of MAC infection in tissues at the time of detectable bacteremia. In the present study, MAC was isolated from the bone marrow of all subjects at the time bacteremia was detected. This difference may be explained by the fact that Torriani et al. [8] used only histologie examination to detect MAC. On the basis of the results reported here, AFB staining is an insensitive method for the detection of low levels of tissue infection (<15,000 cfu/g). In a subsequent autopsy study, Torriani et al. [21] reported a sensitivity of 70% for the detection of AFB by stain, which could account for the 30% of tissues in which AFB was not detected.

In the second autopsy study by Torriani et al. [21], blood and several tissues from 10 MAC-infected subjects who had received macrolide-based combination therapy were quantitatively cultured for MAC. MAC was isolated from the blood and tissues of 7, of whom 6 had had an episode of unresolved or recurrent bacteremia within 3 months of death. Four subjects had discontinued treatment before death (range, 0.7–6.8 months). The correlation between MAC load in blood and tissues was significant (ρ = .88, P = .002). The authors concluded that blood culture is useful as a practical measure of the impact of chronic therapy on infected tissues. The present study evaluated the initial treatment of MAC and did not find such a correlation. As Torriani et al. [21] noted, their findings cannot be applied to the early phase of MAC treatment but only to chronic therapy.

Survival time was significantly associated with baseline bone marrow core MAC load (P = .02). Six of 7 persons with relatively low baseline MAC loads (<20,000 cfu/g) in bone marrow core samples survived ⩽48 weeks after study entry. In contrast, 8 of 9 patients with higher core MAC loads (>39,000 cfu/g) died or withdrew from the study in very poor health within 48 weeks, and most died within 6 months. Untreated MAC infection has been associated with decreased survival in HIV-infected persons [35, 36], and patients receiving MAC prophylaxis have improved survival rates [37, 38]. High MAC load in tissues may be a cause of increased mortality or a noncausally associated factor. High MAC loads may contribute to mortality by enhancing immunodeficiency-related mechanisms that are not reversed by effective MAC treatment. Alternatively, persons with the highest tissue MAC loads may be more immunocom-promised independently of MAC infection, and mortality may remain high despite effective MAC therapy. However, effective treatment of persons with comparatively low tissue MAC loads may interrupt the progression of mechanisms contributing to decreased survival. Improved methods for detection of early tissue infection could also improve therapeutic outcome. The survival analysis is limited by the small sample size and censoring of survival at 1 year. The baseline MAC load in blood did not correlate with survival, although a significant relationship has been reported [39]. It is likely that the relationship between blood MAC load and survival was not detected in the present study because of the small sample size. In contrast, a significant correlation between bone marrow core MAC load and survival was detected. This suggests that there may be a stronger relationship between bone marrow MAC load and survival than between blood MAC load and survival.

HIV-1 loads were high (mean, 5.08 log10 copies/mL) at baseline among the study patients who received no antiretroviral therapy or nucleoside monotherapy. The 95% CI for the rate of change per week in mean log10 HIV-1 RNA level during the first 8 weeks of MAC treatment (−0.03 to 0.05) indicates that the mean HIV load is unlikely to have decreased >0.24 logs or increased by >0.40 logs after 8 weeks, despite significant decreases in MAC loads. Also, the mean HIV load had not significantly decreased 24 weeks after study entry. Recent studies have indicated that mycobacterial infections increase HIV replication [40–42]. In a case-control study of nearly 200 HIV-infected subjects with or without MAC infection, Havlir et al. [43] showed that the onset of MAC bacteremia is associated with a small (0.4 log10 copies/mL) but significant increase in HIV-1 RNA levels. Goletti et al. [44] showed that plasma HIV load increased 5–160-fold during the acute phase of M. tuberculosis disease in HIV-infected patients. Clinical studies also have suggested that effective antimycobacterial therapy may not decrease HIV replication. In studies of African patients with tuberculosis [45, 46], HIV loads increased or did not change despite 6 months of effective tuberculosis treatment.

The incidence of MAC disease in HIV-infected persons has decreased dramatically since the introduction of potent antiretroviral therapy, which may result in restoration of protective immunity against MAC infection. An ongoing study (AIDS Clinical Trials Group Protocol 393) is evaluating the need for MAC maintenance therapy in persons who respond to potent anti-HIV therapy and in whom MAC is likely to have been eradicated from tissues after at least 12 months of macrolide-based treatment as assessed by bone marrow culture. Eradication of MAC from tissues may be possible with antimycobacterial chemotherapy, at least for patients with low initial MAC loads. For example, study patient 6 had no evidence of MAC in any tissues examined at autopsy after the study was completed. The results of this study also indicate that the likelihood and timing of tissue sterilization by MAC treatment cannot be determined from the blood culture response. On the basis of these findings, a bone marrow biopsy for MAC culture is advisable before MAC maintenance therapy is discontinued for persons who respond to potent antiretroviral treatment until more data are available. However, sterilization of bone marrow does not ensure that MAC has been eradicated from all tissues. Lymph nodes, liver, and spleen were found to be more heavily infected than bone marrow in autopsy studies [6, 21].

A more complete understanding of the natural history and pathogenesis of this disease will require an improved knowledge of the specific host immune responses to MAC infection (particularly in tissues), predisposing immune defects, and organism virulence factors. In addition, the complex interplay among HIV infection, MAC disease, and host defenses must be better understood. Clinical trials are underway to further define the crucial immune mechanisms and to determine whether potent antiretroviral therapy restores MAC-specific protective immunity.

Previous SectionNext Section

Division of AIDS Treatment Research Initiative 007 Study Group

In addition to the authors, study group members include the following persons: Maureen Power, Division of AIDS, NIAID; John Pelosi, McKesson BioServices; and Bernard Landry (current affiliation: Medlmmune, Gaithersburg, MD) and Mary Mallory-Smith, Social and Scientific Systems, Rockville, MD; Roy Steigbigel, State University of New York, Stony Brook; Jonathan Cohn, University of Maryland (current affiliation: Wayne State University, Detroit), and Eliot Siegel, VA Medical Center, Baltimore; and James Sampson, Research and Education Group, Portland, OR.

Previous SectionNext Section

Acknowledgments

We thank the patients who volunteered to participate in this study. Microbiology studies were performed by P. Cruz, C. Park, and M. Vasquez at the Non-Tuberculous Mycobacteria Reference Laboratory, Children's Hospital of Los Angeles. Abbott Laboratories supplied the clarithromycin used in this study, and Lederle Laboratories supplied the ethambutol. The clinical sites and investigators of record who conducted this study were Dolores Peterson, University of Texas Southwestern Medical Center, Dallas; Roy Steigbigel, State University of New York, Stony Brook; Jonathan Cohn, University of Maryland, Baltimore; and James Sampson, Research and Education Group, Portland, OR. We also thank the analytic programmers for this study, Eunice Yu and Wendy Watson of Westat, Rockville, MD.

Previous SectionNext Section

Footnotes

  • ↵a Study group members are listed after the text.

  • Informed consent was obtained from all patients participating in the study in accordance with the human experimentation guidelines of the US Department of Health and Human Services and the National Institute of Allergy and Infectious Diseases.

  • Financial support: Division of AIDS Treatment Research Initiative; NIH (AI-15123).

  • Received September 8, 1998.
  • Revision received March 17, 1999.
Previous Section
 

References

  1. 1.↵
    1. Horsburgh C Jr.
    . Mycobacterium avium complex infection in the acquired immunodeficiency syndrome. N Engl J Med 1991;324:1332-8.
    MedlineWeb of Science
  2. 2.
    1. Ellner JJ,
    2. Goldberger MJ,
    3. Parenti DM
    . Mycobacterium avium infection and AIDS: a therapeutic dilemma in rapid evolution. J Infect Dis 1991;163:1326-35.
    Abstract/FREE Full Text
  3. 3.
    1. Young LS
    . Mycobacterium avium complex infection. J Infect Dis 1988;157:863-7.
    Abstract/FREE Full Text
  4. 4.↵
    1. Nightingale SD,
    2. Byrd LT,
    3. Southern PM,
    4. Jockusch JD,
    5. Cal SX,
    6. Wynne BA
    . Incidence of Mycobacterium avium—intracellulare complex bacteremia in human immunodeficiency virus—positive patients. J Infect Dis 1992;165:1082-5.
    Abstract/FREE Full Text
  5. 5.↵
    1. Klatt EC,
    2. Meyer PR
    . Pathology of the spleen in the acquired immunodeficiency syndrome. Arch Pathol Lab Med 1987;111:1050-3.
    MedlineWeb of Science
  6. 6.↵
    1. Wong B,
    2. Edwards FF,
    3. Kiehn TE,
    4. et al
    . Continuous high-grade Mycobacterium avium—intracellulare bacteremia in patients with the acquired immune deficiency syndrome. Am J Med 1985;78:35-40.
    MedlineWeb of Science
  7. 7.↵
    1. Kemper CA,
    2. Havlir D,
    3. Bartok AE,
    4. et al
    . Transient bacteremia due to Mycobacterium avium complex in patients with AIDS. J Infect Dis 1994;170:488-93.
    Abstract/FREE Full Text
  8. 8.↵
    1. Torriani FJ,
    2. McCutchan JA,
    3. Bozzette SA,
    4. Grafe MR,
    5. Havlir DV
    . Autopsy findings in AIDS patients with Mycobacterium avium complex bacteremia. J Infect Dis 1994;170:1601-5.
    Abstract/FREE Full Text
  9. 9.↵
    1. Kemper CA,
    2. Meng TC,
    3. Nussbaum J,
    4. et al
    . Treatment of Mycobacterium avium complex bacteremia in AIDS with a four-drug oral regimen: rifampin, ethambutol, clofazimine, and ciprofloxacin. Ann Intern Med 1992;116:466-72.
    Abstract/FREE Full Text
  10. 10.
    1. Chiu J,
    2. Nussbaum J,
    3. Bozzette S,
    4. et al
    . Treatment of disseminated Mycobacterium avium complex infection in AIDS with amikacin, ethambutol, rifampin, and ciprofloxacin. Ann Intern Med 1990;113:358-61.
    Abstract/FREE Full Text
  11. 11.↵
    1. Hoy J,
    2. Mijch A,
    3. Sandland M,
    4. Grayson L,
    5. Lucas R,
    6. Dwyer B
    . Quadrupledrug therapy for Mycobacterium avium—intracellulare bacteremia in AIDS patients. J Infect Dis 1990;161:801-5.
    Abstract/FREE Full Text
  12. 12.↵
    1. Young LS,
    2. Wiviott L,
    3. Wu M,
    4. Kolonoski P,
    5. Bolan R,
    6. Inderlied CB
    . Azithromycin for treatment of Mycobacterium avium—intracellulare complex infection in patients with AIDS. Lancet 1991;338:1107-9.
    CrossRefMedlineWeb of Science
  13. 13.
    1. Dautzenberg B,
    2. Truffot C,
    3. Legris S,
    4. et al
    . Activity of clarithromycin against Mycobacterium avium infection in patients with the acquired immune deficiency syndrome: a controlled clinical trial. Am Rev Respir Dis 1991;144:564-9.
    MedlineWeb of Science
  14. 14.
    1. Chaisson RE,
    2. Benson CA,
    3. Dubé MP,
    4. et al
    . Clarithromycin therapy for bacteremic Mycobacterium avium complex disease: a randomized, double-blind, dose-ranging study in patients with AIDS. Ann Intern Med 1994;121:905-11.
    Abstract/FREE Full Text
  15. 15.↵
    1. Saint-Marc T,
    2. Marneff E,
    3. Touraine JL
    . Mycobacterium avium intracellulare infections: treatment with a clarithromycin-clofazimine combination: 18 cases. Presse Med 1993;22:1903-7.
    MedlineWeb of Science
  16. 16.↵
    1. Shafran SD,
    2. Singer J,
    3. Zarowny DP,
    4. et al
    . A comparison of two regimens for the treatment of Mycobacterium avium complex bacteremia in AIDS: rifabutin, ethambutol, and clarithromycin versus rifampin, ethambutol, clofazimine, and ciprofloxacin. N Engl J Med 1996;335:377-83.
    CrossRefMedlineWeb of Science
  17. 17.↵
    1. Dubé MP,
    2. Sattler FR,
    3. Torriani FJ,
    4. et al
    . A randomized evaluation of ethambutol for prevention of relapse and drug resistance during treatment of Mycobacterium avium complex bacteremia with clarithromycin-based combination therapy. J Infect Dis 1997;176:1225-32.
    Abstract/FREE Full Text
  18. 18.
    1. May T,
    2. Brel F,
    3. Beuscart C,
    4. et al
    . Comparison of combination therapy regimens for treatment of human immunodeficiency virus—infected patients with disseminated bacteremia due to Mycobacterium avium. Clin Infect Dis 1997;25:621-9.
    Abstract/FREE Full Text
  19. 19.
    1. Chaisson RE,
    2. Keiser P,
    3. Pierce M,
    4. et al
    . Clarithromycin and ethambutol with or without clofazimine for the treatment of bacteremic Mycobacterium avium complex disease in patients with HIV infection. AIDS 1997;11:311-7.
    CrossRefMedlineWeb of Science
  20. 20.↵
    1. Burman WJ,
    2. Stone BL,
    3. Rietmeijer CA,
    4. Maslow J,
    5. Cohn DL,
    6. Reves RR
    . Long term outcomes of Mycobacterium avium complex bacteremia using a clarithromycin-containing regimen. AIDS 1998;12:1309-15.
    MedlineWeb of Science
  21. 21.↵
    1. Torriani FJ,
    2. Behling CA,
    3. McCutchan JA,
    4. Haubrich RH,
    5. Havlir DV
    . Disseminated Mycobacterium avium complex: correlation between blood and tissue burden. J Infect Dis 1996;173:942-9.
    Abstract/FREE Full Text
  22. 22.↵
    1. Benson CA,
    2. Korvick JA
    1. Inderlied CB,
    2. Nash KA
    . Microbiology and in vitro susceptibility testing. In: Benson CA, Korvick JA, editors. Mycobacterium avium-complex infection: progress in research and treatment. New York: Marcel Dekker; 1995. p. 109-40.
  23. 23.↵
    1. Wayne LG,
    2. Sramek HA
    . Agents of newly recognized or infrequently encountered mycobacterial diseases. Clin Microbiol Rev 1992;5:1-25.
    Abstract/FREE Full Text
  24. 24.↵
    1. Lorian V
    1. Inderlied CB,
    2. Nash KA
    . Antimycobacterial agents: in vitro susceptibility testing, spectra of activity, mechanisms of action and resistance, and assays for activity in biologic fluids. In: Lorian V, editor. Antibiotics in laboratory medicine. 4th ed. Baltimore: Williams & Wilkins; 1996. p. 127-75.
  25. 25.↵
    1. Arbeit RD,
    2. Slutsky A,
    3. Barber TW,
    4. et al
    . Genetic diversity among strains of Mycobacterium avium causing monoclonal and polyclonal bacteremia in patients with AIDS. J Infect Dis 1993;167:1384-90.
    Abstract/FREE Full Text
  26. 26.↵
    1. Nash KA,
    2. Inderlied CB
    . Genetic basis of macrolide resistance in Mycobacterium avium isolated from patients with disseminated disease. Antimicrob Agents Chemother 1995;39:2625-30.
    Abstract/FREE Full Text
  27. 27.↵
    1. Wiley EL,
    2. Mulhollan TJ,
    3. Beck B,
    4. Tyndall JA,
    5. Freeman RG
    . Polyclonal antibodies raised against Bacillus Calmette-Guerin, Mycobacterium duvalii, and Mycobacterium paratuberculosis used to detect mycobacteria in tissue with the use of immunohistochemical techniques. Am J Clin Pathol 1990;94:307-12.
    MedlineWeb of Science
  28. 28.↵
    1. Searle S,
    2. Casella G,
    3. McCulloch C
    . Variance components. New York: John Wiley & Sons; 1992. p. 44-105.
  29. 29.↵
    1. Miller R
    . Survival analysis. New York: John Wiley & Sons; 1981. p. 46-63.
  30. 30.↵
    SAS/STAT user's guide. Cary, NC: SAS Institute; 1989.
  31. 31.↵
    1. Havlir D,
    2. Kemper CA,
    3. Deresinski SC
    . Reproducibility of lysis-centrifugation cultures for quantification of Mycobacterium avium complex bacteremia. J Clin Microbiol 1993;31:1794-8.
    Abstract/FREE Full Text
  32. 32.↵
    1. Solis OG,
    2. Belmonte AH,
    3. Ramaswamy G,
    4. Tchertkoff V
    . Pseudogaucher cells in Mycobacterium avium intracellulare infections in acquired immune deficiency syndrome (AIDS). Am J Clin Pathol 1986;85:233-5.
    MedlineWeb of Science
  33. 33.↵
    1. Klatt EC,
    2. Jensen DF,
    3. Meyer PR
    . Pathology of Mycobacterium avium—intracellulare infection in acquired immunodeficiency syndrome. Hum Pathol 1987;18:709-14.
    MedlineWeb of Science
  34. 34.↵
    1. Kilby JM,
    2. Marques MB,
    3. Jaye DL,
    4. Tabereaux PB,
    5. Reddy VB,
    6. Waites KB
    . The yield of bone marrow biopsy and culture compared with blood culture in the evaluation of HIV-infected patients for mycobacterial and fungal infections. Am J Med 1998;104:123-8.
    CrossRefMedlineWeb of Science
  35. 35.↵
    1. Chin DP,
    2. Reingold AL,
    3. Stone EN,
    4. et al
    . The impact of Mycobacterium avium complex bacteremia and its treatment on survival of AIDS patients—a prospective study. J Infect Dis 1994;170:578-84.
    Abstract/FREE Full Text
  36. 36.↵
    1. Horsburgh CR Jr.,
    2. Havlik JA,
    3. Ellis DA,
    4. et al
    . Survival of patients with acquired immune deficiency syndrome and disseminated Mycobacterium avium complex infection with and without antimycobacterial chemotherapy. Am Rev Respir Dis 1991;144:557-9.
    MedlineWeb of Science
  37. 37.↵
    1. Pierce M,
    2. Crampton S,
    3. Henry D,
    4. et al
    . A randomized trial of clarithromycin as prophylaxis against disseminated Mycobacterium avium complex infection in patients with advanced acquired immunodeficiency syndrome. N Engl J Med 1996;335:384-91.
    CrossRefMedlineWeb of Science
  38. 38.↵
    1. Moore RD,
    2. Chaisson RE
    . Survival analysis of two controlled trials of rifabutin prophylaxis against Mycobacterium avium complex in AIDS. AIDS 1995;9:1337-42.
    MedlineWeb of Science
  39. 39.↵
    1. Horsburgh CR Jr.,
    2. Metchock B,
    3. Gordon SM,
    4. Havlik JA Jr.,
    5. McGowan JE Jr.,
    6. Thompson SE III.
    . Predictors of survival in patients with AIDS and disseminated Mycobacterium avium complex disease. J Infect Dis 1994;170:573-7.
    Abstract/FREE Full Text
  40. 40.↵
    1. Nakata K,
    2. Rom WN,
    3. Honda Y,
    4. et al
    . Mycobacterium tuberculosis enhances human immunodeficiency virus-1 replication in the lung. Am J Respir Crit Care Med 1997;155:996-1003.
    Abstract/FREE Full Text
  41. 41.
    1. Orenstein JM,
    2. Fox C,
    3. Wahl SM
    . Macrophages as a source of HIV during opportunistic infections. Science 1997;276:1857-61.
    Abstract/FREE Full Text
  42. 42.↵
    1. Denis M,
    2. Ghadirian E
    . Mycobacterium avium infection in HIV-1—infected subjects increases monokine secretion and is associated with enhanced viral load and diminished immune response to viral antigens. Clin Exp Immunol 1994;97:76-82.
    MedlineWeb of Science
  43. 43.↵
    1. Havlir DV,
    2. Haubrich R,
    3. Hwang J,
    4. et al
    . Human immunodeficiency virus replication in AIDS patients with Mycobacterium avium complex bacteremia: a case control study. J Infect Dis 1998;177:595-9.
    Abstract/FREE Full Text
  44. 44.↵
    1. Goletti D,
    2. Weissman D,
    3. Jackson RW,
    4. et al
    . Effect of Mycobacterium tuberculosis on HIV replication. Role of immune activation. J Immunol 1996;157:1271-8.
    Abstract
  45. 45.↵
    1. Morris L,
    2. Martin DJ,
    3. Sacks L,
    4. et al
    . Program and abstracts of the Fifth Conference on Retroviruses and Opportunistic Infections (Chicago). Alexandria, VA: Foundation for Retrovirology and Human Health; 1998. Persistent elevation of HIV viral load during therapy for tuberculosis [abstract 259]; p. 128.
  46. 46.↵
    1. Wallis RS,
    2. Nsubuga P,
    3. Whalen C,
    4. et al
    . Pentoxifylline therapy in human immunodeficiency virus—seropositive persons with tuberculosis: a randomized, controlled trial. J Infect Dis 1996;174:727-33.
    Abstract/FREE Full Text
| Table of Contents

This Article

  1. J Infect Dis. (1999) 180 (2): 438-447. doi: 10.1086/314865
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 205 (5)
  1. Journal of Infectious Diseases
  1. Alert me to new issues

Most