Comparison of Costs of Strategies for Measuring Levels of Human Immunodeficiency Virus Type 1 RNA in Plasma by Using Amplicor and Ultra Direct Assays

Abstract

The costs of four algorithms for monitoring plasma human immunodeficiency virus type 1 RNA were compared. For patients with strong virologic responses, the use of Ultra Direct exclusively was the cheapest strategy. For patients with weak virologic responses, small savings could be obtained by using Amplicor and retesting only samples with values below 500 copies/ml with Ultra Direct.

The goal of antiretroviral therapy is to reduce and maintain plasma human immunodeficiency virus type 1 (HIV-1) RNA levels (PVL) below 400 to 500 copies/ml (1, 7). Assays capable of quantifying PVL as low as 20 to 40 copies/ml have recently become available (6). However, while earlier assays were able to quantitate up to 750,000 copies/ml, some newer assays are limited up to 65,000 copies/ml. Despite increasing availability of the newer assays, no guidelines have been provided regarding the most cost-effective strategy to implement them. In this paper, we compare the costs of several strategies for monitoring PVL in patients receiving antiretroviral therapy by using data from three clinical studies and a treatment distribution program.

The INCAS (Italy, The Netherlands, Canada, Australia Study) trial compared zidovudine (ZDV) plus nevirapine (NVP), ZDV plus didanosine (ddI), and ZDV plus NVP plus ddI among 150 antiretroviral-naive patients (4). PVL was measured at weeks 0, 1, 2, and 4 and every 4 weeks for 52 weeks. In the second study, 22 patients with CD4 counts below 50 cells/mm³ and with intolerance, toxicity, or disease progression with nucleoside-based antiretroviral therapy received NVP, indinavir (IDV), and lamivudine (3TC) (2). PVL was measured at weeks 0, 1, 2, 4, 6, 8, 12, 16, 20, and 24. In the third study, 48 patients with intolerance to ZDV or disease progression despite ZDV therapy received stavudine (d4T) and 3TC (8). PVL was measured at weeks 0, 2, 4, and 8. The primary results of these three studies have been reported (2, 4, 8) and focus on changes in CD4 counts and PVL over time.

PVL determinations were performed with the Amplicor HIV Monitor assay (Roche Diagnostics, Mississauga, Ontario, Canada) (5). Samples with less than 500 copies/ml were retested with the Ultra Direct assay (Roche Molecular System, Alameda, Calif.) with lower and upper limits of quantitation of 20 and 50,000 copies/ml, respectively.

To compare the costs of strategies for measuring PVL in a clinical setting, we used data from participants in the Drug Treatment Program at the BC Centre for Excellence in HIV/AIDS, a provincewide distribution system of therapies for individuals with HIV or AIDS (3). Since the program’s inception in October 1992, 2925 individuals have received antiretroviral therapy. PVL has been measured quarterly since July 1996 with the Amplicor kit. Values below 500 copies/ml were not retested with Ultra Direct. Many individuals were not antiretroviral naive at the time of their first (baseline) PVL measurement.

Four strategies for measuring PVL were compared. Strategy 1 specified testing of plasma samples with Amplicor and retesting with Ultra Direct if the Amplicor result was below 500 copies/ml. Strategy 2 specified testing with Ultra Direct (Amplicor) if the PVL at the previous visit was less (or more) than 500 copies/ml and retesting with Amplicor (Ultra Direct) if the current result exceeded 50,000 copies/ml (or was below 500 copies/ml). The third strategy specified testing baseline plasma samples with Amplicor, but all later samples with Ultra Direct and retesting with Amplicor if the Ultra Direct result exceeded 50,000 copies/ml. The fourth strategy was to test baseline plasma samples with Amplicor and all other samples with Ultra Direct. The average cost per PVL measurement according to each strategy was calculated retrospectively for each study, assuming the cost of Amplicor was $100 (Canadian) and the cost of Ultra Direct was $100, $120, or $150 (Canadian). The cost of implementing strategy 2 in the drug treatment program was not calculated, since there were not enough longitudinal data to estimate probabilities of a PVL measurement above (or below) 500 copies/ml when it had been below (or above) 500 copies/ml at a previous visit.

While PVL is often measured monthly in clinical trials, it is currently recommended that PVL be measured only quarterly in clinical settings for patients on stable therapy (1). Since strategy 2 depended on prior PVL results, we replicated those analyses, assuming that PVL results would only be available quarterly. This was only done with data from the INCAS trial, because the other studies did not have sufficient follow-up to conduct such analyses.

The costs of the PVL testing strategies are shown by study in Table 1. The ability of a given combination therapy to reduce PVL below the limit of quantitation of the Amplicor kit (500 copies/ml) directly affects the cost per visit of each strategy. Although strategy 2 was the most cost-effective for the ZDV plus NVP and ZDV plus ddI arms of the INCAS trial when monthly PVL measurements were available, the cost per PVL measurement increased by 10 to 14% when only quarterly PVL measurements were available. The cost of implementing strategy 2 depends on the proportion of previous PVL measurements below the limit of quantitation and the degree of concordance between current and past measurements.

TABLE 1.

Average cost per HIV-1 RNA measurement by treatment group, testing strategy, and assumed cost of the Ultra Direct assay

Treatment group (n)	% of PVL		Strategy	Cost of Ultra Direct assaya
	<500 copies/ml	>50,000 copies/ml		$100	$120	$150
INCAS trial
ZDV + NVP (46)	12.6	15.9	1	113	115	119
			2	106	110	115
			3	116	136	166
			4	100	120	150
ZDV + ddI + NVP (51)	66.3	2.0	1	166	179	200
			2	111	126	148
			3	102	122	152
			4	100	120	150
ZDV + ddI (53)	36.3	7.8	1	136	143	154
			2	108	117	129
			3	108	128	158
			4	100	120	150
2nd study; 3TC + IDV + NVP (22)	52.6	5.2	1	153	163	179
			2	113	124	142
			3	105	125	155
			4	100	120	150
3rd study; d4T + 3TC (48)	26.8	15.9	1	127	132	140
			2	116	123	133
			3	116	136	166
			4	100	120	150
Drug Treatment Program (2,925)	29.0	21.0	1	129	135	143
			3	121	141	171
			4	100	120	150

^aAssuming the cost of the Amplicor assay as $100. 

Figure 1 shows a comparison of costs of strategies 1, 3, and 4 in a hypothetical population of patients for whom PVL is measured at months 0, 3, 6, 9, and 12, assuming Amplicor and Ultra Direct cost $100 and $120, respectively. The cost of strategy 4 is less than that of strategy 1 if the proportion of follow-up visits for which PVL is <500 copies/ml exceeds 100× (1 − cost of Amplicor/cost of Ultra Direct). Strategy 2 was not included in this analysis, since we did not have good estimates of the probabilities of PVL being above or below 500 copies/ml given the previous measurement.

FIG. 1.

Comparison of costs of strategies 1, 3, and 4 in a hypothetical situation assuming the costs of Amplicor and Ultra Direct are $100 and $120 and that PVL is measured at months 0, 3, 6, 9, and 12. Strategy 1, test with Amplicor and retest with Ultra Direct if PVL is <500; strategy 3, test with Amplicor at month 0, test follow-up samples with Ultra Direct, and retest with Amplicor if PVL is >50,000; strategy 4, test with Amplicor at month 0, and test all follow-up samples with Ultra Direct.

We have shown that among patients with strong virologic responses, the use of Ultra Direct for all PVL measurements was substantially cheaper than using Amplicor and retesting values below 500 copies/ml with Ultra Direct. Among patients with weak virologic responses, small savings could be obtained by using Amplicor and retesting values below 500 copies/ml with Ultra Direct instead of testing all values with Ultra Direct. If accuracy at the upper end of the PVL scale is not required, using solely the assay with a lower limit of quantitation can result in substantial cost savings. For laboratories with the capability to access patients’ previous PVL measurements, the use of previous PVL measurements to determine which assay to use for the present measurement was found to be cost effective if frequent PVL measurements are available and partially suppressive therapies are used.