A viral load result is usually described in terms of the number of HIV RNA copies per millilitre of blood (copies/ml). In HIV-positive people who have not developed symptoms, a viral load higher than 100,000 is considered to be high, and below 10,000 is considered low.

These numbers are often written in a form known as the logarithmic scale, such as 10⁵ (which is spoken as 'ten to the power five'). This means that the actual number is 10 x 10 x 10 x 10 x 10. Another way of thinking of 10⁵ is as 1 with the decimal point moved five places to the right, which is the same as 1 plus 5 zeros - 100,000. Changes in viral load are also described using this log scale, for example:

• A one log change is a ten-fold change. An example of a one log fall might be 40,000 copies to 4,000 copies. Put another way, this is a 90% fall.

• A two log fall might be for example, 40,000 copies to 400 copies. This is a 99% or hundred-fold fall.

• A three log fall might be for example, 40,000 copies to 40 copies. This is a 99.9% or thousand-fold fall.

Fractions of logs are harder to remember, because they dont correspond to round-number percentages. For example, a 0.5 log fall in viral load is a 66.6% or two-thirds fall in viral load. A 1.5 log fall is approximately a 96% reduction.

So why is this log scale used at all? The answer is that it is a useful way for researchers to present trial data which include a wide range of values, from very high ones to very low ones. Most other people are more familiar with absolute numbers.

Factors affecting viral load

Changes in viral load tend to precede and mirror changes in the CD4 count as the viral load increases over time, the CD4 count decreases although anti-HIV drugs can reverse these changes, at least in the short- to medium-term.

Researchers no longer believe that HIV maintains a stable set point for years. Instead, there is evidence that if HIV is untreated, then viral load steadily increases from the set point over the years. One study analysed blood samples taken over a 17-year period. It found viral load increased by an average of 0.12 log per year, although a higher rate of increase was predictive of more rapid disease progression. Another study looked at eight men in the MACS study and found that viral load increased gradually over several years and that disease progression was associated with a viral load above 100,000 copies.

As with CD4 counts and other laboratory tests, what matters most is the trend in viral load over time, rather than any one single test result. The same viral load test used on the same sample of blood can produce a different result because of the degree of variability in the test. The degree of variation may be anything up to 0.3 log, so two separate tests on the same blood sample might come up with two different results. For example, the difference between 50,000 and 100,000 copies is just 0.3 log, yet this is also the difference between a medium and a high disease progression risk. Similarly, the difference between 25,000 copies and 50,000 copies is also 0.3 log, and this is the difference between low and medium viral load in some guidelines for treatment.

Various factors can cause a temporary blip in the viral load, especially things that stimulate the immune system such as vaccinations. People who develop an opportunistic infection tend to experience a temporary increase in viral load of around 1 log, which usually returns to its previous level within a couple of months of the successful treatment of the infection (although there is some evidence that TB may lead to a permanent increase in viral load).

It is also possible that viral load, like the CD4 count, follows a diurnal rhythm, so that it will always be lower at certain times of the day. Some clinics recommend that patients have blood drawn at a similar time on each visit to avoid this sort of confounding factor.

In all individuals, sex hormones affect T-cell function and the production of immune system messengers. This has led researchers to postulate that womens viral load may vary over the course of a menstrual cycle (the cycle of ovulation and monthly bleeding in women). But there is conflicting evidence about the effect of the menstrual cycle on viral load and CD4 count. A study of 14 HIV-infected women found that the 10 women who ovulated had a small but statistically significant decline in viral load from the pre-ovulation stage, associated with high oestrogen levels, to post-ovulation when progesterone is produced. When data from the women who did not ovulate were added to the analysis, no significant change in viral load was detected (Greenblatt 2000). However, there are a number of problems with the interpretation of this study. Although the difference in viral load between the two groups is statistically significant (i.e. not due to chance), the authors of the study did not address the question of normal variation in viral load results, nor the lack of clinical significance of such a small viral load difference. A more recent study of 55 women found that menstruation did not affect viral load in the blood nor CD4 count but that viral load did vary in the female genital tract during the course of one menstrual cycle (Reichelderfer 2000).

Finally, viral load is also influenced by the conditions in which the blood is stored and the length of time it is kept at room temperature, and last but not least, the potential for human error in carrying out laboratory tests.

For all these reasons, it is important not to make changes on the basis of a single viral load count.

See also Viral load blips in Anti-HIV therapy: Changing treatment for more discussion of the significance of small changes in viral load.

References

Donovan RM et al. Changes in viral load markers during AIDS-associated opportunistic diseases in human immunodeficiency virus-infected persons. Journal of Infectious Diseases 174: 401-403, 1996.

Greenblatt RM et al. Impact of the ovulatory cycle on virologic and immunologic markers in HIV-infected women. Journal of Infectious Diseases 181(1): 82-90, 2000.

Mullins J. Consistent features of HIV-1 evolution in vivo. 38th Interscience Conference on Antimicrobial Agents and Chemotherapy, San Diego, abstract S-100, 1998.

Reichelderfer PS et al. Effect of menstrual cycle on HIV-1levels in the peripheral blood and genital tract. AIDS 14(14): 2101-2107, 2000.

Sabin CA et al. The natural history of viral load over 17 years of HIV infection. Fourth Annual Meeting of the British HIV Association. Oxford, abstract 11, 1998.