Clinical and Resistance Consequences of Misquantification of Plasma and Cerebrospinal Fluid Human Immunodeficiency Virus Type 1 (HIV-1) RNA in Samples from an HIV-1 Subtype G-Infected Patient

Abstract

Human immunodeficiency virus (HIV) load is the main marker used to monitor antiviral treatment efficacy and resistance. We report a case of underquantification of HIV type 1 (HIV-1) RNA in plasma and cerebrospinal fluid from an HIV-1 subtype G-infected woman, leading to delayed diagnosis of HIV encephalitis and to the emergence of drug resistance.

Human immunodeficiency virus (HIV) load is the main marker used to monitor antiretroviral treatment efficacy and resistance (3). Commercial PCR-based HIV type 1 (HIV-1) RNA assays use hybridization primers and probes based on the HIV-1 subtype B genome, which is prevalent in Western countries. However, hybridization problems can arise with non-subtype B HIV-1 genomes. Most commercial PCR assays target highly conserved regions, such as the integrase gene in the Abbott real-time PCR assay and gag in the former Roche COBAS AMPLICOR HIV-1 MONITOR assay version 1.5 and the new Roche Cobas TaqMan real-time quantitative HIV-1 RNA assay. Oliver et al. showed a mean difference of 0.05 log₁₀ RNA copies/ml between the results of the Roche AMPLICOR and TaqMan assays when the tests were applied to 191 samples, whereas Damond et al. reported respective mean values of 4.2 and 2.9 log₁₀ copies/ml for 160 plasma samples (1, 4). Gueudin et al. confirmed that the TaqMan assay gave lower values than the Abbott and AMPLICOR systems (2). Recently, Wirden et al., comparing Abbott and TaqMan assay results for stored samples, reported differences exceeding 0.5 log for about 20% of samples and 1 log for 3% of samples (5). The clinical implications of such discrepancies are controversial.

We report the case of a 35-year-old HIV-1 subtype G-infected Congolese woman living in France for whom misquantification of the viral loads in serial samples led to misdiagnosis, with major neurological complications (Table 1). Highly active antiretroviral therapy (with stavudine, lamivudine, and nelfinavir) was initiated in 2002, when her CD4 cell count was 311/mm³ and her plasma viral load was 441,600 RNA copies/ml according to the AMPLICOR assay. A diffuse CD8⁺ lymphocytosis syndrome was first suspected in 2002; lymphocytic meningoencephalitis was diagnosed in 2004, and an episode of zoster encephalitis followed in January 2006. At the latter date, her plasma viral load was 165,000 RNA copies/ml (by the AMPLICOR assay) and the L90M protease mutation associated with nelfinavir resistance in non-B subtypes was detected. In March 2006, she was switched to abacavir, lamivudine, and ritonavir-boosted atazanavir, and her plasma viral load result from the ultrasensitive AMPLICOR assay fell below 20 copies/ml. During this period, the AMPLICOR assay was replaced by the new Roche TaqMan technology in our laboratory. Her viral load remained undetectable by this new assay. In March 2007, she was admitted to the intensive care unit for refractory seizures. Lymphoma was diagnosed, and HIV encephalitis was ruled out on the basis of her undetectable plasma and cerebrospinal fluid (CSF) HIV loads and the high Epstein-Barr virus load in the same samples. Brain biopsy subsequently ruled out lymphoma, and CD8 vascularitis was diagnosed.

TABLE 1.

Virological and immunological parameters

Sampling date (mo/day/yr)	Antiretroviral treatmenta	CD4 cell count/mm3 (%)b	No. of HIV-1 RNA copies/ml detected by:						Major plasma resistance mutation(s)c	Antiretroviral concns (ng/ml) in:
			AMPLICOR assay		TaqMan assay		Abbott real-time assay
			Plasma	CSF	Plasma	CSF	Plasma	CSF		Plasma	CSF
4/17/02	None	311 (6)	441,660
7/1/02	d4T, 3TC, NFV	249 (8)	<200
2/3/03			<200
7/7/04			87,700				54,667
1/26/06		356 (6)	165,000						RT, no mutation;
7/20/06	ABC, 3TC, ATV/r	517 (8)	<200						PRO, L90M
9/6/06		495 (8)	65,800
11/14/06		174 (9)	<20	<200			<40	440
1/12/07		504 (14)			<40
3/6/07					<40	<400	4,229	14,610		ATV, <30; ABC,	ATV, <30; ABC, <10;
4/10/07	ABC, 3TC, LPV/r	192 (11)			<40	2,060	2,047	268,980		242; 3TC, 625	3TC, 144
4/20/07										LPV, <30; ABC,	LPV, <30; ABC, <10;
6/19/07					<40					94; 3TC, <10	3TC, 240
8/22/07		333 (9)			59
9/21/07		398 (11)			<40		1,293
12/11/07		248 (12)			224		44,856		RT, L74V and M184V;
1/14/08	AZT, TDF, DRV/r						2,541		PRO, I54V and
5/9/08							<40	<40	V82F	DRV, 5,784; TDF, 20; FTC, <10	DRV, <5; TDF, <5; FTC, <10

^ad4T, stavudine; 3TC, lamivudine; NFV, nelfinavir; ABC, abacavir; ATV/r, ritonavir-atazanavir; LPV/r, ritonavir-lopinavir; AZT, zidovudine; TDF, tenofovir; DRV/r, ritonavir-darunavir.

^bPercentage of CD4 cells among total lymphocyte population.

^cMutations included on the 2008 International AIDS Society list (www.iasusa.org/). RT, reverse transcriptase; PRO, protease.

Her worsening clinical condition prompted us to suspect HIV-1 RNA misquantification by the TaqMan assay. Abbott real-time PCR, a method reported to reliably quantify the different HIV-1 group M subtypes, was used retrospectively to verify the viral loads in plasma samples stored at −80°C (Table 1). The results obtained with the Abbott and AMPLICOR assays were similar for samples from July 2004 and November 2006, but major discrepancies (around 3 log) between Abbott and TaqMan results were observed. When the serial samples in which viral loads were undetectable by the TaqMan assay were retested with the Abbott assay, viral loads were found to be not only detectable but rising, due to the emergence of resistant strains. Genotypic analysis of virions from plasma showed mutations (L74V and M184V) associated with abacavir and lamivudine resistance and also mutations (I54V and V82F) conferring resistance to protease inhibitors. Furthermore, the CSF sample with a viral load undetectable by the TaqMan method was found to contain 14,610 RNA copies/ml by the Abbott assay, pointing to a diagnosis of HIV encephalitis.

Retrospective antiretroviral drug concentration measurements showed undetectable protease inhibitor concentrations in plasma and CSF samples, confirming the patient's poor adherence to the treatment regimen.

In conclusion, the problems encountered here with viral load quantification led to a delay in HIV encephalitis diagnosis and to the use of an inappropriate antiretroviral regimen, undermining this patient's chances of recovery. Various guidelines recommend consistent use of the same assay to monitor the HIV load in a given patient and verification of the results with another assay if the CD4 cell count declines despite apparently stable values (6). This case illustrates the need to carefully manage switches from one viral load assay to another, even for patients who have undetectable viral loads.