LETTER

A major challenge for multicenter clinical trials is quality assurance of specimen collection and specimen handling. This is critical for assay reliability and data interpretation. Clinical trials involving human immunodeficiency virus (HIV) studies require large quantities of blood to assess patient safety, as well as immunological, virological, and pharmacological parameters. While some assays require only the cellular portion of blood, others require only the plasma fraction. To minimize the amount of blood required at each visit, it is necessary to consolidate specimens. Many clinical protocols currently require the use of EDTA-anticoagulated whole blood for molecular assays and acid citrate dextrose (ACD)-anticoagulated whole blood for culture assays. The goal of this study was to determine whether one of the two anticoagulants would be better suited to provide samples for all virology assays.

Several studies have evaluated the effect of anticoagulants on HIV type 1 (HIV-1) RNA stability (1, 4, 7-9; J. Moye, L. Mofenson, W. Meyer, R. Nugent, J. Whitehouse, P. Reichelderfer, Z. Qin, and J. Bethel, Abstr. 5th Conf. Retrovir. Opportun. Infections, abstr. 316, 1998). However, not many studies have evaluated the effect of anticoagulants on the ability to culture HIV (3). Moreover, no studies have looked at the effect of ACD and EDTA anticoagulants on the ability to quantitatively culture HIV. Therefore, the NIH/NIAID/DAIDS-sponsored Virology Quality Assessment (VQA) Program initiated a study to investigate this question.

Over a period of 4 months, a total of 88 quantitative microcultures were set up using ACD (n = 44)- and EDTA (n = 44)-anticoagulated blood, collected at 20 separate donations, from 14 HIV-1-positive donors. Cultures were set up within 6 h of collection (n = 40) and again after 24 h, following either an overnight incubation at room temperature (n = 40) or after an overnight shipment (n = 8), using a method previously described (2, 5, 6). This number of cultures was sufficient to give 90% power to differentiate a 0.57 log₁₀ change in infectious units per million cells (IUPM) or a change of one dilution in the culture.

The results from the first set of analyses demonstrated an average loss in IUPM of 0.71 log₁₀ when the blood was processed after 24 h, as compared to blood processed within 6 hours of collection (P < 0.0001). The average loss in IUPM from the ACD cultures was 0.56 log₁₀, and the average loss in IUPM from the EDTA cultures was 0.85 log₁₀; the difference between these losses was not significant (P = 0.16). The baseline titer in ACD-treated blood was only 0.05 log₁₀ lower than the baseline titer for EDTA-treated blood (P = 0.78), and the difference in the mean titer between the anticoagulants, averaged over the two time points, was 0.10 log₁₀ (P = 0.44). In the second part of the analyses, the average loss in culture titer in shipped samples (0.94 log₁₀) was nearly the same as the average loss in samples held overnight in the laboratory (1.0 log₁₀, P = 0.80). Together, these data suggest that EDTA- and ACD-treated blood gave comparable results for quantitative microcultures, regardless of the handling, and that the most deleterious effect on any HIV culture titer was time.

Historically, culture assays have not been done using EDTA-anticoagulated whole blood. Issues surrounding the effect of chelating cations on cellular viability possibly played a role in this decision. However, these new data suggest that EDTA does not have an adverse effect on the ability to quantitatively recover HIV by culture. Moreover, the need to consolidate specimens and minimize blood volumes warrants the use of one anticoagulant for all assays whenever possible. This study supports the concept that protocols could prescribe the use of EDTA-anticoagulated whole blood for all virology assays.