Correspondence
Rifabutin Prophylaxis against Mycobacterium avium Complex Infection
N Engl J Med 1994; 330:436-438February 10, 1994
- Article
To the Editor:
The article on rifabutin prophylaxis against Mycobacterium avium complex disease by Nightingale et al. (Sept. 16 issue)1 provides important information. We are perplexed by one point, however. How, in a randomized trial of 1146 patients, could the group treated with rifabutin have a 14 percent higher mean base-line CD4+ lymphocyte count, the key surrogate marker of the progression of human immunodeficiency virus (HIV) disease (mean CD4+ count, 63.5 cells per cubic millimeter, as compared with 55.5 cells per cubic millimeter for patients in the placebo group)? On the basis of data given in Table 1 of the article, the difference in the CD4+ count between groups is statistically significant (approximate T = 2.5 in both studies combined; two-tailed P = 0.01). Do the authors have any explanation for this difference?
The authors also report that the difference between groups in the incidence of bacteremia remained statistically significant after covariate adjustment for the base-line CD4+ count, but no such information is provided for the more important clinical manifestations, some of which show only borderline statistical significance in unadjusted form, as presented in their Table 3. The base-line CD4+ count needs to be adjusted for the analysis of the data on clinical manifestations.
1 ReferencesRonald H. Goldschmidt, M.D.
Norman Hearst, M.D., M.P.H.
Donald B. Chambers, Ph.D.
University of California, San Francisco, San Francisco, CA 941101
Nightingale SD ,Cameron DW ,Gordin FM , et al. Two controlled trials of rifabutin prophylaxis against Mycobacterium avium complex infection in AIDS. N Engl J Med 1993;329:828-833
Full Text | Web of Science | Medline
To the Editor:
The report on the prevention of M. avium complex infections with rifabutin in patients with AIDS was potentially misleading. The stated goal was to assess the efficacy of primary prophylaxis in patients with CD4+ lymphocyte counts of 200 cells per cubic millimeter or less. However, the mean base-line CD4+ counts in the randomized groups were less than 70 per cubic millimeter. Less than one quarter of patients in the two studies had counts above 100 per cubic millimeter. M. avium complex infections occur infrequently in patients with counts above 100 per cubic millimeter1. In another study, the incidence of such infections increased substantially in patients with CD4+ counts below 60 per cubic millimeter2. These two observations call into question the validity of recommendations that primary prophylaxis of M. avium complex infections with rifabutin be initiated when CD4+ counts are 200 per cubic millimeter or lower.
Furthermore, given an average wholesale cost of more than $6.50 per day, the cost effectiveness of approximately two years of primary prophylaxis should be seriously questioned. In one study in children, HIV-positive patients with effectively treated M. avium complex infections did not have significantly shorter lifespans than patients without such infections3. Nightingale et al. also did not report a significant reduction in mortality with rifabutin prophylaxis.
3 ReferencesArthur S. Zbrozek, R.Ph., M.Sc., M.B.A.
University of Texas Medical Branch, Galveston, TX 775551
Kotloff RM . Infection caused by nontuberculous mycobacteria: clinical aspects. Semin Roentgenol 1993;28:131-138
CrossRef | Web of Science | Medline2
Nightingale SD ,Byrd LT ,Southern PM ,Jockusch JD ,Cal SX ,Wynne BA . Incidence of Mycobacterium avium-intracellulare complex bacteremia in human immunodeficiency virus-positive patients. J Infect Dis 1992;165:1082-1085
CrossRef | Web of Science | Medline3
Lewis LL ,Butler KM ,Husson RN , et al. Defining the population of human immunodeficiency virus-infected children at risk for Mycobacterium avium-intracellulare infection. J Pediatr 1992;121:677-683
CrossRef | Web of Science | Medline
To the Editor:
As prophylaxis is used more widely, more patients with unrecognized disseminated M. avium complex infection and a high bacillary burden may receive single-drug prophylaxis. This approach favors the selection of drug-resistant mutants. Although most patients with disseminated M. avium complex infection are symptomatic, 26 of the 1172 patients (2.2 percent) in the rifabutin trials reported by Nightingale et al. had unsuspected disseminated infection when they were considered for study entry. The selection of rifabutin-resistant mutants has been described during the treatment of two patients with AIDS who had disseminated infection due to M. haemophilum, an organism with antibiotic susceptibilities similar to those of M. avium1.
Thus, there is a possibility that widespread application of prophylaxis may induce rifabutin resistance among strains of M. avium. Additional data should be collected on susceptibility profiles of M. avium isolates from patients who do not respond to prophylaxis, and particular caution should be exercised to avoid single-drug treatment of patients with established disease.
We agree with Masur et al. (Sept. 16 issue)2 that one or more blood cultures for M. avium should be obtained before prophylaxis is begun in patients with clinical or laboratory features suggestive of disseminated disease. A routine mycobacterial blood culture might be especially prudent in patients with CD4+ counts of less than 50 per cubic millimeter.
2 ReferencesC. Fordham von Reyn, M.D.
Dartmouth-Hitchcock Medical Center, Lebanon, NH 03756Sheldon T. Brown, M.D.
Bronx Veterans Affairs Medical Center, Bronx, NY 10468Robert D. Arbeit, M.D.
Boston Veterans Affairs Medical Center, Boston, MA 021301
Bernard EM ,Edwards FF ,Kiehn TE ,Brown ST ,Armstrong D . Activities of antimicrobial agents against clinical isolates of Mycobacterium haemophilum. Antimicrob Agents Chemother 1993;37:2323-2326
Web of Science | Medline2
Masur H, Public Health Service Task Force on Prophylaxis and Therapy for Mycobacterium avium complex
Full Text | Web of Science | Medline
To the Editor:
The report by Nightingale et al. and the Special Report by Masur et al. present good evidence of the efficacy of rifabutin in preventing disseminated M. avium complex disease in patients with AIDS, although the data do not show significant prolongation of life. However, this treatment, now virtually certain to achieve wide use in patients with AIDS and those with HIV infection (presumably limited only by patient compliance), is likely to widen the current epidemic of multidrug-resistant tuberculosis.
It is universally accepted that multidrug-resistant tuberculosis is caused by patient irregularity in taking prescribed medication or physician difficulty or inexperience in prescribing appropriate medication. Both these avenues lead to monotherapy, which in the presence of a large bacterial population directly selects for drug-resistant mutants.
Herein lies the problem.
Nontypical infiltrates are well described in patients with tuberculosis who are coinfected with HIV. The misdiagnosis of tuberculosis in these patients is still a problem. With widespread rifabutin use, some persons with unsuspected active tuberculosis and large mycobacterial populations will be treated with M. avium complex prophylaxis under the assumption that any pulmonary infiltrate must be due to pneumocystis or bacterial pneumonia.
Rifabutin has cross-resistance with rifampin, the most important antituberculosis drug. Nightingale et al. do mention that “it is therefore imperative that active tuberculosis be excluded in an HIV-infected patient before rifabutin monotherapy is begun.” Masur et al. suggest a chest film and tuberculin test to exclude the possibility of active disease.
The exclusion of the possibility of active tuberculosis is such an important step that it should be incorporated into the product labeling of rifabutin for this indication, especially since many physicians may be unfamiliar with what the process requires1. At a minimum, the process of exclusion requires a good posterior-anterior chest film and three smears for acid-fast bacilli, read by a competent observer. Even in patients with a clear chest film, and especially in the presence of any infiltrate, particularly in patients with known or past Pneumocystis carinii pneumonia or treated tuberculosis, rifabutin should not be prescribed until the results of the smears are available. A tuberculin test in patients with low CD4+ cell counts is of limited benefit2.
Rifabutin prophylaxis against M. avium complex may benefit many patients with AIDS, but its judicious, careful use will certainly benefit society at large.
2 ReferencesLee B. Reichman, M.D., M.P.H.
Reynard J. McDonald, M.D.
Bonita T. Mangura, M.D.
New Jersey Medical School, National Tuberculosis Center, Newark, NJ 07103-24061
Flora GS ,Modilevsky T ,Antoniskis D ,Barnes PF . Undiagnosed tuberculosis in patients with human immunodeficiency virus infection. Chest 1990;98:1056-1059
CrossRef | Web of Science | Medline2
Markowitz N ,Hansen NI ,Wilcosky TC , et al. Tuberculin and anergy testing in HIV-seropositive and HIV-seronegative persons. Ann Intern Med 1993;119:185-193
Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Unlike Goldschmidt et al. we were relieved, rather than perplexed, to find that the mean CD4+ counts of the rifabutin and placebo groups differed by only 8 per cubic millimeter. As we noted in our paper, the subjects were randomized in blocks of four according to site, to preclude a regional or institutional bias in treatment allocation, rather than according to the CD4+ count. The calculated significance of the observed difference between mean base-line CD4+ counts is a function of sample size (the more subjects, the smaller the standard error), but its clinical importance is another matter. With the use of a formula derived elsewhere,1 the expected one-year incidence of M. avium bacteremia would be 13.3 percent in patients with CD4+ counts of 63.5 per cubic millimeter and 14.6 percent in those with CD4+ counts of 55.5 per cubic millimeter. The goal of randomization was to make the study groups comparable rather than identical; we believe this goal was achieved.
Zbrozek misstates the goal of our study, which was to assess the efficacy of primary prophylaxis in patients with AIDS (i.e., those with an AIDS-defining opportunistic event according to the 1987 criteria of the Centers for Disease Control and Prevention2) and CD4+ counts of 200 per cubic millimeter or less. M. avium complex infections are infrequent in patients who have not previously had some other AIDS-defining opportunistic event3. Such patients were excluded from our study and from our recommendations based on it. We agree that the costs and benefits of prophylaxis against M. avium complex infection deserve additional discussion, and we are among those who have done so elsewhere4. Regarding the effect of prophylaxis on mortality, we reiterate that patients in the placebo group as well as those in the rifabutin group were screened monthly for M. avium complex bacteremia, even when asymptomatic, and were offered all available therapies as soon as bacteremia was detected. We believe this is why the survival advantage of rifabutin treatment over placebo (33 vs. 47 deaths during the blinded phase of the trial) was not greater.
von Reyn et al. note that 2.2 percent of our study patients had unsuspected M. avium bacteremia at base line. It is uncertain whether such latently infected patients have the high bacillary burden that favors the selection of drug-resistant mutants5. We encourage others to monitor rifabutin-treated populations for the emergence of drug-resistant mycobacteria, but we did not observe this ourselves.
In response to Reichman et al.: we believe that a chest film and a blood culture are often useful components of the clinical evaluation of patients being considered for rifabutin prophylaxis, but we would not ordinarily obtain sputum if the chest films revealed no abnormalities. In urban areas such as the one where they practice, the effectiveness of public health measures will be far more important in determining the future course of the epidemic of multidrug-resistant tuberculosis than rifabutin use.
5 ReferencesStephen D. Nightingale, M.D.
University of Texas Southwestern Medical Center, Dallas, TX 75235D. William Cameron, M.D.
Ottawa General Hospital, Ottawa, ON K1H 8L6, Canadafor the Rifabutin Study Group Investigators
1
Nightingale SD ,Jockusch JD ,Haslund I ,Cal SX ,Peterson DM ,Loss SD . Logarithmic relationship of the CD4 count to survival in patients with human immunodeficiency virus infection. Arch Intern Med 1993;153:1313-1318
CrossRef | Web of Science | Medline2
Revision of the CDC surveillance case definition for acquired immunodeficiency syndromeMMWR Morb Mortal Wkly Rep 1987;36:Suppl 1:1S-15S
3
Horsburgh CR Jr ,Selik RM . The epidemiology of disseminated nontuberculous mycobacterial infection in the acquired immunodeficiency syndrome (AIDS). Am Rev Respir Dis 1989;139:4-7
CrossRef | Web of Science | Medline4
Nightingale SD . CD4 thresholds for prophylaxis against Mycobacterium avium-intracellulare complex infection and other opportunistic events in HIV-infected patients. Med Decis Making 1993;13:394-394 abstract.5
Torriani F, McCutchan JA, Grafe M, et al. Syndrome of sustained Mycobacterium avium complex (MAC) bacteremia without tissue invasion in AIDS patients. In: Program and abstracts of the 33rd Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, October 17-20, 1993. Washington, D.C.: American Society for Microbiology, 1993:376. abstract.
