Infections are diagnosed by detecting the presence of specific antibodies, since they are found in the serum in virtually 100% of infected people. With the aim of minimising the risk of false-negative results, all the techniques are highly sensitive, with close to 99% accuracy. It is difficult to achieve 100% sensitivity as seroconversion does not take place until 3-4 weeks after infection and, moreover, infected people can be seronegative as a result of immunity defects. The increase in sensitivity entails a reduction in specificity, which with the current techniques is around 99%. Furthermore, a lower prevalence of HIV infection in the studied population reduces the positive predictive value, meaning that the likelihood of false-positive results occurring with low HIV infection rates is higher. All positive results must therefore be confirmed by a confirmatory test [1].

The most commonly-used confirmatory techniques are Western Blot (WB) and recombinant immunoblot or line immunoassay (LIA). They have at least the same sensitivity as ELISA and superior specificity. Both techniques can incorporate HIV-2 envelope antigens, thereby making it possible to diagnose HIV-2.

The methodology for both methods has now been semi-automated, making them easier to perform. However, the results can be subjective as the reading is based on observation of the presence of coloured bands which correspond to different viral proteins [7]. Each laboratory must therefore establish a strict system for taking reactivity readings.

The detection of specific antibodies indicates exposure to the virus and infection and direct detection of the viral p24 antigen introduces a dynamic concept to serology; since it is a marker of virus replication, it provides information on the current status of the infection, it is detected in initial stages of the infection or in the development of AIDS, and it serves as back-up to the serological diagnosis in situations where the detection of antibodies is inconclusive [8]. The p24 antigen can be measured in serum and plasma with uptake ELISA techniques that increase sensitivity, which at present can be as high as 8 pg/ml.

The main characteristics of the available systems for screening and confirmation of HIV infection and the extent to which they are automated are shown below. Table 1 shows the methodology, generation, format and main characteristics of fully automated systems for HIV diagnosis. Table 2 lists the main features of ELISA semi-automated systems. Table 3 shows the confirmatory assays and their features.

Viral load, or plasma viraemia, is the amount of virus circulating in plasma and refers to the number of copies of HIV RNA present per millilitre, expressed as copies/ml or log₁₀. The HIV viral load and CD4 lymphocyte count are currently the two most used prognostic markers of clinical progression of the HIV infection [9]; the first is an indicator of virus replication and the second reflects how the host’s immune system is working. Both parameters are monitored every 3-6 months, although some authors [10] are now advocating annual monitoring, at least of the CD4 count. A high initial viral load is a marker of rapid progression and clinical progression is always preceded by an increase in the viral load. Changes in plasma HIV-1 RNA levels are the best marker for response to antiretroviral therapy, since they are sensitive, fast and reliable in terms of good management of the patient with HIV infection. The success of antiretroviral therapy is defined in the majority of guidelines as the suppression of plasma viraemia and clinically, it is considered as such when the patient has two successive viral loads below 50 copies/ml [11], although this cut-off point has been the subject of debate in the last year [12-14].

Viral load should not be used routinely to diagnose the infection, except in special cases such as confirmation of neonatal infection.

Given how important it is to measure this parameter, accurate, reproducible, cost-effective techniques are required [15-17].

There are currently a number of different platforms available for quantifying HIV viral load:

• Abbott RealTime HIV-1 m2000rt (Abbott Molecular Diagnostics): combines the extraction in the automated m2000sp system and qrtPCR amplification and detection with the automated m2000rt equipment.
• COBAS AmpliPrep TaqMan HIV-1 48 (Roche Molecular Diagnostics): uses the COBAS AmpliPrep for the extraction of the sample and the COBAS TaqMan 48 v2.0 for the qrtPCR amplification and detection.
• NucliSens EasyQ HIV-1 v1.2 (bioMérieux): uses the EasyMAG for the extraction and the EasyQ for qrtPCR amplification and detection.
• VERSANT HIV-1 RNA 1.0 (kPCR) (Siemens): this system combines a sample extraction module (SP module) with an amplification and detection module (AD module).
• Artus HIVirus-1 QIAsymphony SP/AS and Rotor-Gene Q (Qiagen): the extraction is carried out in the QIAsymphonySP/AS equipment and the amplification and detection in the Rotor-Gene Q.

All of these use real-time PCR to quantify HIV RNA in plasma, displacing previously-used techniques such as conventional PCR, the NASBA technology, ligase chain reaction or signal amplification. The advantages of real-time PCR are its high levels of analytical sensitivity, reproducibility and linearity, dynamic range and the fact that the current techniques can quantify different types and subtypes of HIV, not only HIV-1 subtype B. These instruments incorporate automated RNA extraction methods which significantly reduce the time it takes to obtain results and the risk of contamination, and a larger number of samples can be processed. They all start with the reverse transcription of HIV-1 RNA to cDNA. The target is then amplified with real-time PCR, generating fluorescent molecules which are ligated to oligonucleotide probes which specifically bind to the amplified product. When the fluorescence exceeds a minimum signal, the number of PCR cycles (Ct) that have taken place is used for the quantification. One of the advantages of this technology is that it enables detection and quantification of the target without having to handle the reaction tubes. The main features of these assays are shown in Tables 4-7.

There is intrinsic variability between the different assays and it is therefore recommended that the monitoring of a patient’s viral load is always done with the same system; if a change of assay is unavoidable, it must be ensured that a new baseline quantification of the HIV-1 RNA is obtained with the technique to be used thereafter. For quantification in the different subtypes of HIV, nowadays, all these techniques have been adapted to detect a large number of non-B subtypes.

Drug resistance testing has become a key component of proper HIV clinical care and is currently recommended by most HIV treatment guidelines [11, 18, 19] as the standard of care, both in terms of choosing the most effective antiretroviral therapy for first-line regimens, by investigating the transmission of drug-resistant variants at the time of diagnoses (primary resistance) and in terms of the selection of active drugs in subsequent failures to first- or further-treatment-line failures (acquired resistance).

Although resistance assays can be categorized as either phenotyping or genotyping, the genotyping assays are the most widely-adopted in routine diagnostic laboratories. Phenotyping resistance assays measure the extent to which an antiretroviral drug inhibits virus replication in vitro and is usually performed by demonstrating a change in the inhibitory concentration (IC) that is required to inhibit in vitro growth by 50 percent (IC50) compared with virus replication in the absence of drug. Results are reported as a fold change in drug susceptibility of the patient sample in relation to a wild type reference strain. These assays are currently performed using Recombinant Virus Assay technology. They are typically available from commercial laboratories and are rarely performed in routine diagnostic laboratories. An overview of the phenotyping methods and their advantages and limitations can be found in Garcia et al. [4].

Genotyping resistance assays detect the presence of specific drug-resistance mutations. Mutations are then interpreted into a drug-resistance report by using resistance interpretation systems. The report typically gives a classification into “susceptible", "possibly resistant" or "resistant" for each antiretroviral agent. Interpretation rules and algorithms can vary greatly from one to another [20]. Currently the protease (Pro), reverse transcriptase (RT), integrase (INT) and the V3 loop are the regions of interest for HIV-1 DNA sequencing. Pro, RT and INT are used to investigate resistance to protease, reverse transcriptase (nucleoside/otide and non-nucleoside analogues) and integrase inhibitors, while V3 sequencing is used for co-receptor usage estimation (tropism testing) prior to the use of CCR5 antagonists.

The most widely-used tests in routine clinical laboratories are those investigating RT and Pro mutations. Two commercially available FDA-approved and CE-marked kits are available: TRUGENE HIV-1 Genotyping Assay (Siemens NAD) and ViroSeq^® HIV-1 Genotyping System (Celera-Abbott diagnostics). Although there are some differences in performance, both tests are labour intensive, both for the sequence generation and detection steps. Both methods generally provide concordant results in the way they detect mutations, but they differ in the way resistance reports are given [21].

Compared to phenotyping, the advantages of genotyping tests include less complexity, shorter turn-around time and lower cost, while their main limitations are: being an indirect measure of resistance; that they only interrogate known mutations; the difficulty in interpretation when complex patterns of mutations are present; the differential effect on resistance that combinations of individual mutations may have compared to individual mutations alone; and that they cannot detect low-abundance resistant mutations (minor variants).

Table 8 shows the main features of two tests, including technical, complexity, biosafety and cost-effectiveness aspects.

Conventional resistance tests can detect mutations which comprise over 20% of all the viral genomes amplified with RT-PCR. We are now at the stage where it is becoming important to detect mutations that might be between 1% and 20%, since it has been demonstrated that these patients are at greatest risk of failing therapy [22]. Techniques are therefore being introduced which are capable of detecting and/or quantifying mutation levels below the 20% threshold imposed by the currently-available techniques.

One of the methods already available, allele-specific real-time PCR (AS-PCR), can detect mutations present in as low as 0.01% of the population as a whole [23]. Despite this extraordinary sensitivity, however, the great disadvantage of this method is that an assay for every allele of every mutation is needed. This has meant that the only results presented have been for detecting the main mutations associated with resistance to non-nucleoside analogues (K103N, Y181C, G190A) and M184V [23-25], a mutation that has been associated with resistance to the main nucleoside analogues.

This drawback is overcome by the massively ultra deep parallel sequencing techniques (UDPS) which have been developed in the last few years and are able to detect mutations from 0.5%. Moreover, as these are single genome sequencing techniques, they can provide data on the percentage of each mutation compared to the whole population and on the mutational load, which is proving to be extremely useful as a true parameter in the interpretation of resistances. Another advantage of these platforms is that by sequencing single genomes, we are not limited to the known mutations, but with the introduction of the new antiretrovirals, we will be able to read each of the potential positions of interest in the future.

There are currently various UDPS platforms available. Table 9 provides a summary of the different characteristics of the platforms which are more advanced in terms of development, highlighting two principal aspects. The first is the length of the readings generated which, with the Roche platform (454) [26, 27], is close to that obtained by the Sanger method, making them ideal for single genome sequencing, although the great disadvantage is that they provide poor resolution in homopolymeric sequences. In contrast, Solexa [28] and SOLiD [29] have the great advantage over pyrosequencing of providing good resolution of the homopolymeric regions, but are not capable of generating readings beyond 75 bases and they cannot be used for de novo sequencing. Another important aspect is that they are able to sequence the DNA without it first having to be cloned, saving a lot of work and laboratory space. With these new sequencers, the cost of each nucleotide has fallen from $10 in 1990 to $0.01 in 2012.

The great disadvantage of these techniques is that they are not automated. Although systems are being developed to enable the different steps to be automated, in the meantime, it has to be said that they are very laborious manual techniques, they take up a lot of time and require highly qualified personnel. So far, these platforms are not being routinely used in the laboratory. However, we are sure that in the not-too-distant future, that will change and these ultra-sequencing techniques will be available for more general use.

With regard to tropism testing, tropism can be inferred using phenotyping methods (Antivirogram^® (Virco) [30], Pheno Sense^® (Virologic) [31], Phenoscript^® (VIRAlliance) [32] and MT-2 assay [33]), which are only possible in specialised laboratories or private companies that provide the logistics for collection and transport of samples. It can also be determined by using a genotyping method consisting of sequencing the V3 region of the gp140 envelope gene; depending on the sequence the quasi HIV species present, the virus will have R5 tropism (use the co-receptor CCR5) or X4 tropism (use the co-receptor CXCR4). The interest in this determination is that, before being able to use a CCR5 antagonist (the only one approved so far is Maraviroc) it has to be confirmed that the patients are carriers of virus with R5 tropism. At present, there are no commercially-available tests with CE marking or FDA approval for determining viral tropism. These techniques have not been automated in the laboratory and manual procedures have to be used in order to obtain results [34].