Introduction

Accurate estimates of HIV-1 incidence are needed to evaluate the state of the HIV epidemic, to calibrate and validate models of epidemiology, and to plan and assess the impact of prevention interventions [1], [2]. Incidence testing provides a measure of the recent level of HIV-1 transmission in a population [2]. To obtain HIV-1 incidence estimates from cross-sectional cohorts, testing methods have been developed that take advantage of the biological differences between recently-infected and chronically-infected individuals [3], [4]. Three assays used to estimate the proportion of recently-infected individuals in a population are the BED capture enzyme immuno assay (BED-CEIA) [5], the Vironostika-less sensitive assay (V-LS) [6], and assays based on antibody avidity [7], [8], [9]. However, it is not known how serum/plasma sample handling conditions impact data generated by these incidence assays. Because incidence assays are most often performed using stored samples that may have been freeze-thawed for other investigations, it is critical to understand how prior storage and manipulation of samples might affect the data generated by cross-sectional incidence assays. We evaluated the impact of sample storage (4°C, 25°C and 37°C for up to two weeks) and freeze-thaw cycles (up to 15 cycles) on results obtained with the BED-CEIA, V-LS, and avidity assays.

Methods

Results

A total of 428, 381 and 376 samples from the original 31 chronically infected patients were tested using BED-CEIA, V-LS, and avidity assays, respectively. Total samples available for each test differed because of the difference in the initial sample volumes available. For the BED-CEIA, the median normalized optical density (OD-n) of the untreated samples was 3.05 (interquartile range [IQR]: 2.26, 3.22). For the V-LS assay, a median SOD value for untreated samples was 4.54 (IQR: 3.67, 5.19). Samples tested using the avidity assay had a baseline median percent avidity (AI) of 100% (IQR: 78.7%, 100.0%).

Results from the BED-CEIA, V-LS, and avidity assays are shown in Table 1 and Figures S1, S2, S3, S4, S5, S6, S7, S8 and S9. The impact of freeze thaw cycles on testing is shown in Table 2 and Figures S10, S11, and S12. With the exception of conditions in which samples were stored at 37°C, there were no significant differences in results obtained using the control vs. manipulated samples. For samples stored at 37°C a decrease in testing values was observed using the BED-CEIA and V-LS assays. The mean decrease in OD-n for the BED-CEIA was 0.19 for 3 days, 0.34 for 7 days and 0.54 for 15 days. A similar trend was seen with the V-LS assay as the SOD values changed when the samples were incubated at 37°C prior to testing, with a mean decrease of 0.21 for 3 days, 0.71 for 7 days and 0.96 for 15 days. For the avidity assay, a nonsignificant decrease in AI was seen for samples incubated at 37°C prior to testing, with a mean decrease 1.41 for 3 days, 7.71 for 7 days and 1.99 for 15 days.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Mean Change in Assay Results by Handling Condition.

https://doi.org/10.1371/journal.pone.0025899.t001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Mean Change in Assay Results by Number of Freeze Thaw Cycles Prior to Testing.

https://doi.org/10.1371/journal.pone.0025899.t002

Discussion

Our analysis demonstrates that results generated using the BED-CEIA, V-LS, and avidity assays are not significantly affected by most sample-handing conditions. Freezing and thawing samples up to 15 times had no impact on the values generated by any of the three assays. Additionally, storing samples at 4°C or 25°C prior to testing had no statistically significant effect on the test results in the bivariate analyses. We did demonstrate a minor decrease in the values obtained using the BED-CEIA, V-LS, and avidity assays when samples were stored for an extended period at 37°C. Because relatively few aliquots were tested in the evaluation of the 15-day incubation and 15 freeze-thaw cycle conditions, the moderate changes observed in assay test results under those conditions were not conclusive. A larger sample size might provide stronger statistical support. These results are similar to those reported previously for the stability of samples used for testing with HIV-1 antibody assays for the diagnosis of HIV-1 infection [12]. In that study, repeated freeze-thaws and incubation at −20°C and 4°C did not alter the capacity of ELISA and Western blot assays to detect antibodies to HIV. A small decline in antibody reactivity was noted when samples were incubated at 25°C or 37°C for more than 21 days; however, that change did not alter the interpretation of the test results. [12]. Our study extends the previous study by evaluating the effect of freeze-thaws and incubation at different temperatures on the performance of assays that measure different characteristics of the anti-HIV antibody response: the proportion of antibody that is HIV-specific, antibody titer, and antibody avidity. In summary, these data indicate that stored samples previously frozen and thawed multiple times can be used for antibody-based incidence testing. Unless samples were stored at excessive temperatures for an extended period, similar results were obtained using samples subjected to a variety of freeze-thaw and incubation conditions.

Supporting Information

Acknowledgments

Samples and data from Baltimore were supplied by Jeanne Keruly and Dr. Richard Moore. We would like to thank Elaine McKinney for technical assistance.