Mycoplasma incognitus: NEWLY DISCOVERED TREATABLE OPPORTUNISTIC INFECTION?

Researchers at the U. S. Armed Forces Institute of Pathology (AFIP) in Washington, D. C., and the Warren Grant Magnuson Clinical Center at the National Institutes of Health, have found compelling evidence that a previously unrecognized opportunistic infection -- one potentially treatable with antibiotics -- may be a major cause of illness in people with AIDS. Many infections of organs including the brain, spleen, liver, or lymph nodes -- as well as some systemic infections -- might be caused by the newly-discovered organism, called Mycoplasma incognitus. Until now, these infections would be counted among the many which cannot be diagnosed.

While the first report of the organism now known as Mycoplasma incognitus was published over three years ago, most of what is now known was learned later and published last year. And only in the last few weeks has the AIDS research community paid serious attention. Until recently the new organism was mistakenly believed to be a virus, and its discovery seemed to have little immediate relevance to treatment.

Then a series of five articles by Shyh-Ching Lo and others in the American Journal of Tropical Medicine and Hygiene, between February and November 1989, showed:

(1) The new organism is a mycoplasma -- which is potentially treatable. Mycoplasma, a form of life between bacteria and viruses in complexity, was discovered about 100 years ago. Some species are known to cause human diseases.

The published articles only hint that the new organism might be treatable with antibiotics. But scientists at AFIP tested 15 common antibiotics against the Mycoplasma incognitus in the laboratory. A detailed report is being prepared for publication, but because of the public-health importance of the information, AFIP released a list of the drugs and their effective concentrations in a separate document. Doxycycline, tetracycline, clindamycin, lincomycin, and ciprofloxacin were found to be effective against Mycoplasma incognitus. But erythromycin, the antibiotic most commonly used to treat mycoplasma infections, was not effective -- and penicillin, streptomycin, gentamicin, and others also had no effect.

(2) Mycoplasma incognitus was found in the thymus, liver, spleen, lymph node, or brain of 22 of 34 persons who had died of AIDS. The patients who were selected for this autopsy study had all had evidence of organ failures.

(3) In a separate study with different patients, the mycoplasma was found in seven of ten persons with AIDS. Also, a much earlier study had found Mycoplasma incognitus in blood lymphocytes of 12 of 23 living persons with AIDS -- but in none of 22 healthy blood donors used as controls.

(4) The mycoplasma was also found in six HIV-negative patients (with no sign of AIDS) from different parts of the world, who had died in one to seven weeks of an undiagnosed infection.

No one knows how the organism spreads, but evidently it is not by casual contact, as family members of infected persons have not become infected themselves.

(5) Four monkeys were injected with Mycoplasma incognitus; all died in seven to nine months. The organism was found in the spleens of all the monkeys, and in some other organs as well. It was not found in a fifth monkey tested as a control.

(6) Extensive evidence from electron-microscope examinations, from specially designed PCR tests to look for the DNA of Mycoplasma incognitus, and from immunologic tests, showed that the organism was concentrated in lesions in affected organs. Mycoplasma incognitus is unusual in that it often infects and kills tissue without causing an inflammatory reaction, suggesting that it disables or evades part of the immune system.

The publication of this evidence, much of it in November 1989, led to a meeting between Dr. Anthony Fauci, director of NIAID (the National Institute of Allergy and Infectious Diseases) and other AIDS experts, with Dr. Lo and his colleagues at AFIP. The meeting, on December 14, 1989 in San Antonio, was chaired by Dr. Joel B. Baseman, chairman of the Department of Microbiology at the University of Texas Health Sciences Center in San Antonio, an expert on mycoplasma. An article in THE WASHINGTON POST (January 5) quoted Dr. Baseman as saying that Lo's mycoplasma "might be a significant agent for many infectious diseases, not just AIDS. There is enough information to say that this agent is real." The same article quoted Dr. Fauci as saying that Mycoplasma incognitus "may be an important opportunistic infection ...If it's real, it could have an important impact on how doctors look at AIDS patients with unexplained problems."

An in-depth history of the discovery of Mycoplasma incognitus and its early dismissal by parts of the scientific community was published in THE NEW YORK TIMES, January 16, 1990.

What Should Be Done Now?

Awareness of the new importance of Mycoplasma incognitus has not yet spread far in the medical community. The biggest problem is that there is no readily available test for the organism; at this time, there may be only one research lab which can do the test reliably. Other mycoplasmologists are becoming involved, however, and a much easier blood test is being developed. In addition, clinical trials are now being planned.

The AIDS physician and patient community should help make sure that trials proceed quickly. There may also be immediate uses of the new information, for example:

(1) Empirical Use of Antibiotics

Several months ago, before Mycoplasma incognitus was known, Dr. Nathaniel Pier mentioned that he had found good results trying doxycycline for patients who had an apparent infection which could not be diagnosed. (Doxycycline is the antibiotic most often discussed as a possible treatment for Mycoplasma incognitus; however neither it nor any other antibiotic has yet been tested for treating this infection in humans.) Incidentally, the next physician we asked about empirical use of antibiotics preferred erythromycin, which would not be effective against the mycoplasma.

The discovery of Mycoplasma incognitus provides an additional rationale for trying doxycycline (or one of the other antibiotics found effective against this organism in the laboratory) for certain patients, such as those with signs of undiagnosed infection, especially in the central nervous system, spleen, or certain other organs. Patients should know that antibiotics can cause side effects -- some of which, such as overgrowth of Candida, might be more severe in persons with HIV. Physicians and patients must use appropriate precautions.

Physicians' groups could make a major contribution by developing interim guidelines for empirical use of antibiotics -- guidelines to use now before a blood test for Mycoplasma incognitus is available. It may seem premature to develop guidelines with no diagnostic test and no trial results. But physicians are already using antibiotics empirically to treat undiagnosed infections in persons with AIDS or HIV. Instead of waiting months or years for definitive information, it would be helpful to collect reports of such clinical experience with the drugs known to be active against the organism, detailed information about the symptoms of persons who have already been found to be infected, and any special precautions necessary when using the antibiotics to treat persons with AIDS or HIV.

(2) A Rapid Clinical Trial

Research Design

The usual kind of clinical trial -- randomly treating some patients and leaving others untreated, then waiting to see which group does better -- would create some unusual ethical, practical, and scientific problems in this case. Therefore we are suggesting a possible alternative, which research groups, including community-based research organizations, might want to consider. Readers not interested in clinical trial design might want to skip this section. We included it because some of the ideas are not obvious, and we believe they might contribute to discussions of clinical trials.

The scientific problem with the common trial design -- selecting patients who test positive for Mycoplasma incognitus and randomly treating part of the group with a broad-spectrum antibiotic like doxycycline -- is that AIDS/HIV patients have many undiagnosed infections, so patients might improve even if mycoplasma is not what is being treated. We propose the following design which not only overcomes this problem, but also answers the ethical concerns about leaving patients untreated, and at the same time makes a trial much less costly and easier to administer. This design would be especially suitable for community-based research. And it should prove definitively whether or not a given test for mycoplasma is relevant to the decision to use or not to use a particular treatment.

Before beginning the trial, it would be necessary to negotiate arrangements to get blood samples tested for mycoplasma. At this time, AFIP may be the only organization which could do this work.

When the trial begins, the physicians involved would continue their usual treatment, making decisions strictly for their patients' benefit, not for research. Those patients who met clinical criteria suggesting possible Mycoplasma incognitus infection could be enrolled in the trial. A blood sample would be taken before antibiotic treatment began, and sent to the lab for the mycoplasma test. Later, the physicians would rate how well the patients had responded to the treatment.

The research organization would "blind" both parties. The physicians would not know which patients' blood samples had tested positive for mycoplasma. And the lab which tested the samples would not know which patients had responded to the treatment.

The results would be analyzed by a two-by-two table, or similar statistical technique. One dimension of the table would divide patients who did or did not respond to the antibiotic; the other dimension would be whether or not the patients tested positive for mycoplasma. A statistically significant relationship between the two variables would prove that it is possible to use the blood-test information to help predict who is likely to respond to the therapy. Such a result would strongly support the hypothesis that the mycoplasma was causing human disease, although academic scientists could still argue that the mycoplasma might only be a marker for some other organism which was the real target of the treatment. (Even if this possibility were true it would have no practical effect, as it would not diminish the usefulness of the test for predicting which patients would respond to the treatment.)

Note that the trial we propose would not prove that treatment was better than no treatment for any particular group of patients, since no one is randomly assigned to be deliberately left untreated. In theory, this question of whether treatment or no treatment is best is the key issue which trials are designed to address. But in the situation here, after a successful trial as outlined above, that question would not really be the point at issue. For the patients would clearly have needed treatment and have a poor prognosis without it; they would have responded to a generally safe and very well known treatment, and the trial would have generated strong evidence that Mycoplasma incognitus did cause human disease (in addition to the strong pathology evidence already available). The issue of whether the patients would have been better left untreated would only be academic.

One apparent problem with this proposed design is that it would be difficult to calculate in advance the number of patients needed (to prove statistically that the test for Mycoplasma incognitus would be useful for guiding treatment decisions). If the two-by-two table turned out to be strongly unbalanced on either of its dimensions -- for example, if 99 percent of the clinically-selected patients tested positive for the mycoplasma, or if 99 percent tested negative -- the number of patients required for proof would be very large. But again this problem is more apparent than real. For if the table is highly unbalanced, then the result of the study would not only be more expensive to obtain, it would also be less valuable, as it would affect few treatment decisions. Therefore if the first group of patients tested were to show that either dimension of the table was strongly unbalanced, the study could be aborted at that point and redesigned. Therefore, a worst-case maximum number of patients (and cost estimate) could be prepared in advance for planning purposes.

Most observational studies -- those which simply record results of patients' treatment, without randomly assigning patients to receive different regimens -- cannot produce statistical proof of any hypothesis. We believe that the design proposed above is an exception. It avoids the usual constraints by asking a different question: not "does the drug work," but "can information from the test help determine which patients will or will not respond to the drug." By addressing this question instead of the usual one, the proposed trial allows treatment to proceed totally in the patients' interest, without any compromise for research purposes -- while at the same time rigorously proving (or failing to prove, or even disproving) a medically relevant hypothesis.

Summary

The organism previously called a "virus-like infectious agent," discovered by Dr. Shyh-Ching Lo and colleagues at the Armed Forces Institute of Pathology, has been found to be a mycoplasma which is susceptible to several common antibiotics. Even before a blood test is widely available and clinical trials have been done, physicians may want to consider this new information when choosing empirical antibiotic treatment for patients with certain undiagnosed problems. We will report further as more information becomes available.

References

Altman LK. Unusual microbe, once dismissed, is now taken more seriously. THE NEW YORK TIMES, January 16, 1990, page B6.

Booth, W; Specter, M. Microbe may play role in AIDS, other diseases. THE WASHINGTON POST, January 5, 1990, page A3.

Lo SC; Dawson MS; Wong DM; Newton PB 3d; Sonoda MA; Engler WF; Wang RY; Shih JW; Alter JH; Wear DJ. Identification of Mycoplasma incognitus infection in patients with AIDS: an immunohistochemical, in situ hybridization and ultrastructural study. AMERICAN JOURNAL OF TROPICAL MEDICINE AND HYGIENE, November 1989, volume 41, number 5, pages 601- 616.

Lo SC; Shih JW; Newton PB 3d; Wong DM; Hayes MM; Benish JR; Wear DJ; Wang RY. Virus-like infectious agent (VLIA) is a novel pathogenic mycoplasma: Mycoplasma incognitus. AMERICAN JOURNAL OF TROPICAL MEDICINE AND HYGIENE, November 1989, volume 41, number 5, pages 586-600.

Lo SC; Dawson MS; Newton PB 3rd; Sonoda MA; Shih JW; Engler WF; Wang RY; Wear DJ. Association of the virus-like infectious agent originally reported in patients with AIDS with acute fatal disease in previously healthy non-AIDS patients. AMERICAN JOURNAL OF TROPICAL MEDICINE AND HYGIENE, September 1989, volume 41,
number 3, pages 364-376.

Lo SC; Wang RY; Newton PB 3d; Yang NY; Sonoda MA; Shih JW. Fatal infection of silvered leaf monkeys with a virus-like infectious agent (VLIA) derived from a patient with AIDS. AMERICAN JOURNAL OF TROPICAL MEDICINE AND HYGIENE, April 1989, volume 40, number 4, pages 399-409.

Lo SC; Shih JW; Yang NY; Ou CY; Wang RY. A novel virus-like infectious agent in patients with AIDS. AMERICAN JOURNAL OF TROPICAL MEDICINE AND HYGIENE, February 1989, volume 40, number 2, pages 213-226.

In addition, the NEW YORK NATIVE has published frequent and sometimes controversial coverage of this research.