Counting CD4 cells cheaply: diagnostic accuracy study

HIV infection is prevalent in sub-Saharan Africa, but resources for testing and treatment are limited. According to guidelines from the World Health Organization, if CD4 counting is available adults and children over 5 years old should start antiretroviral treatment as soon as their CD4 count drops below 200 cells/micro l, regardless of their clinical stage. Flow cytometric CD4 counting is the best technique for measuring cell counts. However, despite efforts to bring the cost of consumables down for developing countries, the technique is still too expensive for most African communities to consider it.

In Africa a programme to deliver free antiretrovirals has been started, but because the cost of CD4 testing often precludes its use, many patients are allocated to treatment on the basis of clinical criteria rather than CD4 counts. Clinical staging of HIV disease does not fully predict immunological status, so a CD4 cell count is the most effective indicator for starting and assessing immunological response to antiretroviral treatment. At present treatment may not always be targeted to people most likely to benefit. The authors of this study wanted to see if there was a way to make flow cytometry affordable for more African centres so that antiretroviral treatment resources might be better allocated.

How did the researchers tackle the problem?

The authors wanted to test a cheaper flow cytometric method that reduces costs without compromising accuracy. They took advantage of technological advances to develop a flow cytometry method that "miniaturises" the process and uses less reagent but is still accurate. They called this the Blantyre count method (named after the Malawian district where the study took place). The study first compared the Blantyre count technique with two existing validated methods (TruCount and FACSCount) to assess the diagnostic accuracy of the new method. Further study evaluated the effects on clinical decision making.

What did they measure?

The authors looked at three separate sets of blood samples to assess different things. Firstly, 134 consecutive samples sent to the laboratory for HIV testing during two periods in 2006 were included in the CD4 counting comparison study. Cell counts for these were measured using all three methods (one new method and two established methods). Secondly, 253 samples from new patients with HIV during two periods in 2006 were tested for CD4 count using the Blantyre method, and this was compared with clinical staging (performed by staff blind to the CD4 results).

Thirdly, a small study of 28 samples compared the percentage of CD4 cells (out of all lymphocytes) generated by the Blantyre technique (which gives a percentage and an absolute count) and a cheaper variant of the technique (which gives only a percentage count). This was performed because treatment in children under 5 years is begun based on the percentage of CD4 cells in the total lymphocyte count rather than on the absolute CD4 count.

Routinely, papers refer to "the outcome of interest" when they tell us about what they measured. In this study the outcome of interest was not the CD4 or percentage itself but the accuracy of the new technique. So the researchers were primarily interested in the degree of agreement among the various methods of CD4 counting used, and between CD4 flow cytometry results and clinical evaluation.

When performing the cell counts, more than one researcher "read" the tests in each instance. Requiring agreement between two or more readers of the results of tests reduces the likelihood of observational errors and observer bias and makes for more robust research.

What did they find?

The authors found that they needed 20 microl of blood and 10 micro l of counting beads to maintain test quality. At those concentrations the reagent costs per assay were $0.44 (0.21 pound; 0.31 euro) for an absolute and a percentage CD4 count, $0.40 for an absolute count only, and $0.11 for a percentage count only. (Previous initiatives to reduce the price of reagents for CD4 tests in poor countries achieved prices of $3-$6 a test.) The authors report that the cost reduction is "11-fold" for reagents, compared with standard tests.

Importantly, agreement among the Blantyre method and the two existing validated tests was good to excellent for all variables studied. So the authors concluded that they had developed a more affordable method for flow cytometry CD4 counting by refining current techniques without having to sacrifice accuracy. Agreement on CD4 percentage, in particular, was also good between the Blantyre count method and its cheaper variant.

When CD4 counts were compared with clinical staging of disease the study found that CD4 counts varied considerably among patients clinically determined to be in a particular stage, confirming previous observations that clinical staging is not well correlated with cell counts.

Was it a good study?

The costs involved in accurately staging HIV disease need to be reduced so that treatment can be better allocated. The authors of this study are to be commended on their endeavour. It is also good to test a procedure intended for use in an under-resourced setting in that setting to establish its suitability. Even without the need to reduce costs, there is a continuing need for evaluation of and improvement in critical diagnostic tools.

The study might have been too small to detect differences between test results adequately, and, considering that the samples came from just one region in Malawi, the patients may have represented a too similar, or homogeneous, cohort. Follow-up research is needed to see whether this test will be as accurate in other populations.

The study ought to have included treated patients because CD4 counts are increasingly being used to define outcomes and to direct changes in treatment so the performance characteristics in treated patients are important.

The authors themselves also point out that this new approach to CD4 counting by flow cytometry still requires access to expensive machines that require maintenance. Such considerable up-front investment lessens the overall economic benefits of this test, in which money is saved on reagents, and it limits the number of centres that would be able to provide the test.

The cost of CD4 counting can be reduced in two ways to make it more widely available. One is the method that has been reported in this paper and the other is to develop alternative counting methods. One of the rapid responses to this paper, published on bmj.com, said that researchers should not be fixated on reducing the cost of flow cytometry because other methods can also estimate CD4 cell numbers without requiring the same level of technical expertise as flow cytometry. Although the Blantyre method is a cheaper, simplified method of flow cytometry, performing the test still requires a level of technical expertise that may be lacking in the developing world.

However, as the authors of a linked editorial published in the BMJ remarked, "Although the cost of flow cytometry is initially high, strategically located facilities such as regional centres for antiretroviral treatment can provide the high volume of patients needed to offset the capital investment, as long as reagents are affordable. Secondly, there is no dearth of talent in resource constrained settings; only a lack of political will to make the necessary investments in training and quality control." In other words, people can always be trained and services can usually be better structured to improve outcomes.

Competing interests: None declared.

Provenance and peer review: Commissioned; not externally peer reviewed.

1 Erlewyn-Lajeunesse M. CD4 counts and HIV therapy: flow cytometry is not the answer. [electronic response to MacLennan CA et al. Diagnostic accuracy and clinical utility of a simplified low cost method of counting CD4 cells with flow cytometry in Malawi: diagnostic accuracy study.] BMJ 2007. www.bmj.com/cgi/eletters/335/7612/190#173688

2 Madhivanan P, Krupp K. Technological challenges in diagnosis and management of HIV infection in resource limited settings. BMJ 2007;335:165-6.

Kirsten Patrick, former Roger Robinson editorial registrar, BMJ
Email: kpatrick@bmj.com


Student BMJ 2007;15:427-470 ISSN 0966-6494 | December

   

 Printable version    Download PDF    E-mail this to a friend    Respond to this article