NEW VIRAL MEASUREMENT FOR RAPID DRUG TESTING

On December 14, 1989, The New England Journal of Medicine published an editorial and two articles on measuring the amount of HIV in blood. Since then we have heard that many AIDS experts consider this work among the most important of the year. At first it was not clear to us why a new blood test should be so important, when what we need are treatments.

The new tests are important because they tell researchers quickly and accurately which antivirals are working in actual human use -- and how well they are working. It should now take weeks, instead of two years or more, to obtain statistical proof that a treatment does have efficacy in patients.

The new tests could break the logjam of promising drugs waiting for clinical trials. Also, they should make it possible to find the best doses quickly, to develop drug combinations rationally, and even to tell individuals whether a particular treatment regimen is working for them, and when it has stopped working and should be changed.

In the past, the only rapid test for antiviral activity was in a laboratory dish. Even if the drug worked and was known to be safe, many questions remained. Would enough of the drug be absorbed? Would it get to the virus in the body? Could high enough concentrations be achieved? Would the drug remain in the body long enough to have the desired effect? The new tests bypass these questions by directly measuring whether or not patients' viral levels have decreased. They quickly measure antiviral effectiveness of treatments in actual use, not of chemicals in test tubes.

The biggest bottleneck to making new drugs available is the time taken by very slow, cumbersome, and expensive clinical trials. For example, since hundreds of volunteers are needed to prove that each antiviral works, each trial must run at many different sites, requiring time-consuming coordination among institutions in distant cities; for example, dozens of different IRBs (institutional review boards) must approve any change in the protocol of a major trial. By contrast, the new tests can show a clear, unambiguous antiviral effect in a handful of patients in a few weeks.

There are several different approaches to quantifying HIV levels in blood plasma or blood cells. The ones described last month in The New England Journal of Medicine have the most visibility and professional momentum at this time; they are the ones ready for use now. We will focus on them in this article, but also mention alternative approaches.

Background: Viral Cultures

Most (although not all) methods for quantifying HIV in blood use viral cultures; in other words, they determine if the virus is present by attempting to grow it in the laboratory. HIV cultures have been used for several years, but until now they have not proved very useful for measuring the effects of antivirals. The fundamental problem is that cultures remained positive even after treatment, and therefore they did not show how well the treatment worked. No drug yet known can kill all the AIDS virus in patients; therefore, to evaluate a drug today, we need a quantitative test -- one that shows how much virus is present, not just whether or not any is there.

In addition, there are practical difficulties with using viral cultures -- and these remain today. The tests are more expensive than routine blood work like T-cell counts, and they require special laboratory facilities to protect lab workers from exposure to virus. Also, it is difficult to get consistent results with viral cultures -- partly because the tests must use human blood cells as "food" for HIV, and since everyone's cells are different, it is difficult or impossible to reproduce a test exactly. Fortunately the ACTG (AIDS Clinical Trials Group) of the U. S. National Institute of Allergy and Infectious Diseases has certified certain labs as able to perform viral cultures competently. At this time, because of the difficulty of doing viral cultures, the best way to get them done for clinical trials is probably by collaboration between researchers doing the trials and those developing testing methods. The latter usually need ore blood samples for their research -- especially samples collected under the controlled, well-documented conditions of clinical trials.

Measuring AIDS Virus in the Blood

The most important new development for faster testing of new treatments was the publication last month of a means for measuring the amount of HIV in blood (Ho and others, 1989). Two different tests were reported: one for blood plasma, and the other for certain blood cells.

The basic idea of the new tests is so simple that it is surprising no one tried it before. The sample of blood plasma (or blood cells) is successively diluted to lower and lower concentrations, until finally the solution is so dilute that usually there is no virus left in the small amount of the solution which is tested. Ordinary viral cultures are run at all the different dilutions, and the researchers merely note the amount of dilution beyond which the cultures stop being positive. From this information, the amount of virus in the original blood serum or cell sample can be estimated.

Since it is not known how many viral particles may be needed to make a culture become positive, the test does not estimate the actual number of particles, but rather gives results in units called "tissue-culture-infective doses" (TCID) per milliliter of blood. For example, patients with asymptomatic HIV infection were found to have 30 TCID per milliliter of blood plasma -- meaning that one thirtieth of a milliliter was enough to infect a viral culture and make it positive, while less than that amount was not enough.

Dr. Ho and his team found virus in the plasma and blood cells of every one of 54 HIV-positive persons who were not receiving antiviral treatment -- and from none of 22 HIV-negative subjects used as controls. While persons who were asymptomatic had only 30 TCID per millimeter of plasma, persons with AIDS and ARC had over 100 times as much virus, 3500 and 3200 TCID respectively. Blood cells from persons with AIDS or ARC were also found to be more than 100 times as infective as cells from asymptomatic seropositives (20 TCID per million cells for asymptomatic patients, 2200 for AIDS, and 2700 for ARC).

How well do these tests show whether a treatment is working? Seven patients were tested before and after four weeks of AZT; the amount of virus in plasma dropped 94 percent, although unfortunately the level of virus in blood cells did not change greatly. In contrast, in four persons with ARC who were not receiving antiviral treatment, the viral levels in plasma remained stable during the 12 to 20 weeks they were tested.

These results clearly showed antiviral effectiveness of AZT in human use, in a four-week test with only seven patients. In six of the seven, plasma viral levels decreased more than ten times; in the seventh, they decreased almost ten times. But in clinically stable, untreated patients, the levels stayed the same. This complete separation between treated patients and controls suggests that statistical proof of antiviral efficacy could be obtained with a small number of volunteers; and the short time required (four weeks to prove efficacy of AZT, for example) might allow placebo trials to be ethically conducted in clinically stable patients who agreed to risk being untreated for that time (some patients cannot tolerate standard treatment anyway). Placebo trials can usually get results much faster than trials which use an active control such as AZT.

A well-designed, focused program of small, rapid clinical trials could screen the most attractive of the many safe, available, but so far unproven treatment possibilities, and provide numerical measures showing which ones do or do not work to reduce the AIDS virus in actual use by patients. Such a program could quickly and inexpensively test ddI, ddC, compound Q, hypericin, very low doses of AZT, etc., as well as drug combinations, or even proposed diet or lifestyle changes. Some immune modulators might also be found to reduce viral levels indirectly, by helping the body control HIV, even if there is no direct antiviral effect. Such a testing program could give much better guidance for making treatment decisions than any information now available, and greatly reduce the time now taken for clinical trials
and new-drug approval.

(Note: Part II of this article will examine other new evidence that plasma viral level is the best marker available to show the stage of HIV infection. It will also discuss other approaches to measuring HIV in the blood, including R-HEV and quantitative PCR.)

References

Ho, DD and others. Quantitation of Human Immunodeficiency Virus Type 1 in the Blood of Infected Persons. The New England Journal of Medicine, volume 321, number 24, pages 1621-1625, December 14, 1989.