III. HIV THERAPIES

11. ANTIRETROVIRAL THERAPY

Kenny CW CHAN

Today, the standard therapy for HIV/AIDS is highly active antiretroviral therapy (HAART) which comprises three or more drugs of more than one class. Life long treatment is necessary. As the patient often receives additional treatment for opportunistic infections and for complications of antiretroviral therapy itself, it is imperative that the prescribing physician become very familiar with the adverse effects, pharmacokinetic properties and drug interaction profile of all antiretrovirals. The ease with which HIV develops and archives resistance mandates very careful prescription and only by physicians experienced and knowledgeable in its use. Failing this, the current scenario with antibiotics and multi-drug resistant bacteria will be quickly replicated in HIV therapy.

Four classes of 18 antiretroviral agents are currently in use in Hong Kong. This armamentarium will certainly expand quickly. The major features of antiretrovirals are summarised in Box 11.1. The four classes are: (a) Nucleoside reverse transcriptase inhibitors (NRTI); (b) Non-nucleoside reverse transcriptase inhibitors (NNRTI); (c) Protease inhibitors (PI); (d) Entry inhibitor. Currently, co-receptor antagonists and integrase inhibitors are candidate drugs already in clinical trials. New NNRTI and PI with retained efficacy against resistant viruses are also being developed.

Box 11.1

Rationale of antiretroviral therapy

HIV causes progressive damage to the immune system largely through loss of CD4+ T lymphocytes. Immune activation in the chronic phase allows continued HIV entry and replication. As many as 1010 virions are produced and cleared every day. The half-life of circulating virions is estimated to be 1-2 hours.

Antiretroviral therapy interrupts crucial steps of the HIV life cycle. Certain combinations of antiretrovirals are so potent that plasma viral load is suppressed to undetectable levels by conventional assays. This has led to the concept of HAART, which literally means using very potent antiretroviral regimen to control HIV disease. In operational terms, therapy is HAART if sustained suppression of viral load to undetectability is achieved.1 As a result of viral suppression, immune damage is halted and in most cases reversed so that the CD4 count rises and immune capability is reconstituted. The virologic and immunologic benefits have been translated into decreased morbidity and mortality. With HAART, management of HIV disease has taken resemblance to other chronic medical conditions such as diabetes.

It is imperative that only HAART is given for treatment of established HIV infection. By durably reducing viral load to undetectable levels, HAART prevents the development of resistance. Suboptimal therapy such as single or double therapy may have short-term immunologic benefits but will allow detectable viral replication and generation of resistance mutations. Substituting one or even two drugs in a failing regimen may amount to sequential monotherapy and should be avoided.

Despite a vastly improved prognosis, HAART as we know today does not eradicate HIV. Even if viral replication is entirely terminated, a very low level of viraemia is still present, though undetectable with conventional viral load assays. This is the virus released from a pool of latently infected CD4 cells. If therapy is interrupted, virus replication will immediately rekindle. Importantly, this latently infected pool of cells represent a complete archive of all resistance ever happened in the individual, by virtue of the fact that replicating virus continuously enters into this population. Compartmentalisation may be important too. As some antiretrovirals fail to penetrate certain sanctuary sites, there is opportunity for continued viral divergence and replication in certain compartments, such as the genital tract and central nervous system.2

For all these reasons, the central dictum of antiretroviral therapy is that it should be continuous and potent. To this end, adherence of the patient will have to be rigorous. Compliance less than 95% is associated with clinical failure.3 Nevertheless, expected non-adherence is NOT contraindication to treatment. Instead, it should alert the physician to all applicable proactive measures of improving adherence. It is common knowledge that physician's prediction of adherence is poor.

Initiation of therapy (Algorithm 11(A))

The symptomatic patient - HIV-associated symptoms are indications for HAART. However, the urgency varies according to clinical circumstances. A relatively high CD4 count and mild symptoms, e.g. those corresponding to the CDC 'B' category, will allow time for full evaluation and counselling of the patient before optimal treatment is designed and initiated. This will enhance patient adherence in the future. On the other hand, certain opportunistic infections are potentially fatal unless significant immune reconstitution occurs. In these cases, HAART will be more urgently needed. Of note, in those with tuberculosis and a CD4 count >100/μL, treatment should normally be delayed until two months into antituberculous treatment to minimise the occurrence of immune reconstitution disease,4 the latter is described in another chapter (Chapter 15).

The asymptomatic patient - The first antiretroviral regimen carries the best chance of success. Its timing is therefore crucial. In those with symptomatic disease, the cut-off CD4 threshold for treatment initiation is subject to debate. Currently, most authorities suggest that treatment should be initiated with the CD4 count in the range of 200 to 350/μL. This recommendation is based on a multitude of studies comparing mortality based on treatment initiation at various CD4 thresholds. While mortality rises in those with treatment started at a CD4 below 200/μL, it does not significantly differ between those started between 200-350/μL and those started above 350/μL. Within the range between 200 and 350/μL, initiation will depend on the patient's motivation, viral load, and the rate of CD4 decline. These are guidelines meant for reference, as all clinical decisions should be individualised. The pros and cons of early (CD4 >350/μL) and deferred (CD4 200-350/μL) should be communicated clearly to patients to facilitate an informed decision (Box 11.2). In Hong Kong, this issue is complicated by the finding that local Chinese5,6 have a lower natural CD4 level than Caucasians. Hence the cut-off CD4 values recommended for starting ART in some western countries may be overaggressive in Chinese.7

Box 11.2

In general, because of the multiple factors affecting CD4 count and viral load, they should be repeated before any treatment-related decision based on these values is made.

It is controversial if treatment should be started in those who are diagnosed at the stage of seroconversion, and for how long. Although certain disease markers improve with immediate treatment, the clinical impact remains to be fully evaluated. These patients should preferably be referred for enrolment in clinical trials. Nevertheless, in those patients with severe symptoms or failure to rebound from a low CD4 count, HAART will still be indicated according to the usual criteria.

Choosing the appropriate antiretrovirals for the treatment-naïve patient

Once the decision to initiate treatment is made, an appropriate regimen is designed after evaluating the following:

Virus characteristics

Patient characteristics

Drug characteristics

'Preparing' the patient and adherence

Recommended regimens

The current formula of HAART for the treatment-naïve is 2 NRTI + 1 PI or 1 NNRTI. This is based on the general potency of such a regimen. Experience in the last decade however has accumulated to favour certain particular combinations.

PI-containing regimen - There is extensive track record of the PI-based HAART. However, all PIs are not equal in terms of potency and adverse effects. For example, nelfinavir (NFV) is probably the least potent although it is also the most tolerable. Indinavir (IDV) is potent but fraught with multiple adverse effects. Kaletra is the recommended PI for a HAART regimen, based on superior clinical effectiveness which may be due to an optimal balance among potency, patient acceptance, adverse effect and a high barrier to resistance. This recommendation may change as new PIs and formulations of existing PIs become available.

Nevertheless, if Kaletra is not considered the best, alternative PI may be considered based on individual patient profile. The current consensus is that PI should be given with ritonavir (RTV) as the pharmacokinetic booster. RTV has the additional property of inhibiting the P-glycoprotein system and enhancing the cellular penetration of PIs which are substrates of the system.10 More popular combinations are RTV 100-200 mg/IDV 800 mg bid, RTV 100 mg/fosamprenavir (FPV) 700 mg bid, RTV 100 mg/atazanavir 300 mg qd, and RTV 100 mg/SQV1000 mg bid. In fact, Kaletra itself is lopinavir (LPV) coformulated with RTV (LPVr). LPV is not available alone.

NNRTI-containing regimen - In the INCAS study,11 the regimen of ZDV/ddI/NVP demonstrated its superiority over a double-nucleoside regimen. Another prospective trial showed that EFV-based HAART surpassed an equivalent IDV-containing regimen.12 Thus was established the potency and durability of a regimen composing of 2 NRTIs + 1 NNRTI. However, it must be cautioned that NNRTI has a low genetic barrier, with one mutation being sufficient to generate a high level of cross-class resistance. Viral fitness is minimally impaired with these mutations. Therefore, such therapy should preferably be given to those who are motivated and not on medications that may have pharmacokinetic interactions. For all these shortcomings, NNRTI-based treatment is rapidly gaining popularity because it is dosed less frequently and is more tolerable to take. NVP is associated with a more favourable lipid profile than EFV, but is also more prone to serious adverse effects.13 In general, EFV is the recommended NNRTI to use, but treatment decisions should be highly individualised.

The NRTI backbone - The old teaching of using almost any combination of two NRTI as the backbone, with one of them being a thymidine analogue, is now replaced with specific recommendations: 3TC or FTC should be combined with TDF, ABC (abacavir) or ZDV. There is some evidence that ZDV is outperformed by TDF and ABC.14,15 The co-formulation pills of Kivexa (ABC/3TC), Truvada (TDF/FTC) and Combivir (ZDV/3TC) are favoured where applicable, as they simplify dosing and encourage adherence. ddI and d4T are falling out of favour because of significant mitochondrial toxicity. In fact, the two should not be given together because of a high risk of lactic acidosis and neuropathy. Similarly, d4T + ZDV is contraindicated because of antagonism. TDF + ddI has a negative impact on both immunologic and virologic response and should be used with caution. For those who are having toxicities with these suboptimal NRTI backbones, consideration may be given to changing to the recommended combinations. In lipodystrophy, such switching has been shown to be associated with partial reversal of fat atrophy.

Alternative first line regimen - triple NRTI

In limited circumstances where there is strong patient preference or unique drug interaction profile to avoid, a triple combination of NRTIs may be considered. This refers almost exclusively to the combination of AZT/3TC/ABC (Trizivir). Studies have demonstrated an acceptable potency, but durability is doubtful. This regimen is inferior to EFV-containing regimens but probably comparable to NFV- and IDV- based HAART, especially in the presence of VL <105/mL. Triple-NRTI has the advantages of saving at least two classes of antiretrovirals for future use, a low pill burden (two pills a day for the coformulation pill, Trizivir), and avoidance of many significant drug interactions.

On the other hand, triple combinations of most other NRTIs should be avoided. Specifically, the following regimens fare unfavourably in clinical trials: TDF/3TC/ABC, TDF/ddI/3TC, and d4T/ddI/ABC. Many failures of these combinations are associated with K65R which causes resistance to most NRTIs.

Novel regimens

Studies have been conducted with novel regimens, in the hope of finding alternatives that have distinct advantages over the current formula of HAART.

NRTI-sparing treatment - although some encouraging data have been reported with pilot studies of the use of NVP or EFV plus Kaletra,16 an early randomised trial of IDV plus EFV showed inferiority to Combivir plus EFV. In a recent, large trial comparing conventional regimens vs NNRTI+PI, inferiority is again shown with the NRTI-sparing regimens.17 Although NRTI-sparing regimens are not appropriate for the naïve patients, they will be necessary in those with, for example, lactic acidosis secondary to NRTI, extensive resistance to NRTI, and severe lipoatrophy affecting adherence. Studies are under way to identify effective HAART regimens without nucleosides.

Monotherapy - for monotherapy to succeed as HAART, the agent has to be very potent and tolerable. Kaletra monotherapy has been studied in a small trial with acceptable 48-week results.18 However, in all likelihood, a mono-HAART does not exist with the currently available drugs. Furthermore, the availability of combination pills has simplified dosing to such an extent that Kaletra monotherapy has limited use in the real world. In fact, it should be avoided.

Induction-maintenance therapy - the concept of using HAART as induction followed by less potent regimens such as double nucleosides is attractive, but is not supported by clinical evidence. Clinical failure is almost inevitable. It is however possible that induction with quadruple therapy followed by conventional HAART as maintenance may be useful in patients with very high viral loads. Studies to date have been inconclusive.

Structured treatment interruption (STI) - although preliminary data were encouraging, STI with scheduled on and off periods to reduce the total dose administered failed in most large studies. Therefore, it cannot be recommended. Another approach of STI by using the CD4 level to guide the interruption of treatment is also inferior to continuous HAART and should be considered only in exceptional circumstances.

Goals of therapy

Currently, HAART comprising two NRTI and one PI /NNRTI is the preferred regimen. In clinical practice, the goals of HAART are three-folds:

Durable suppression of viral load

Durable suppression of viral load is the desired outcome. Although partial suppression may also be beneficial in immunologic and clinical terms, this is not acceptable as a treatment goal unless no reasonably effective HAART can be constructed, e.g. in the case of multiple resistance. Viral load is the concentration of RNA level in plasma measured by reverse transcription polymerase chain reaction (RT-PCR), branched DNA (bDNA) or Nucleic Acid Sequence Based Amplification (NASBA). It is preferable to follow a patient with the same testing method as results may differ. Patients on HAART should attain suppression of their plasma viral load to below the limit of quantitation (undetectable level). Conventional assays have a detection limit of 400 to 500 copies/mL, whereas ultrasensitive assays have limits of 40-50 copies/mL. Both are acceptable in clinical practice.

Immune reconstitution

In general, immune reconstitution follows the use of HAART and successful suppression of viral load. There is a biphasic increase of CD4 T lymphocytes, where the first rapid wave of increase is represented by mainly memory (CD45RO+) cells, followed a few months later by the addition of naïve (CD45RA+) cells. Proliferative lymphocyte responses to recall antigens and mitogens are increased over time. There is also reduced T cell activation due to reduced viral replication and the T cell receptor repertoire is partially restored. However, immune response specific to HIV antigens remains weak.

Clinically, immune reconstitution is evidenced by the spontaneous remission or improvement of opportunistic infections, e.g. microsporidiosis, and Kaposi's sarcoma. In many conditions, primary and secondary prophylaxis can be safely stopped if there is good immune recovery. On the other hand, immune reconstitution syndrome may also occur. It may be defined as presentation or clinical deterioration of opportunistic infections in HIV infected patients as a direct result of the enhancement of immune responses to these pathogens during HAART.

Reduction of morbidity and mortality

Although immune reconstitution may lead to aggravation of known or subclinical infections in the short term, the overall impact on mortality and morbidity has been one of dramatic reduction. In Hong Kong, median survival after AIDS increased from 29.8 months during the pre-HAART era to >70 months during the HAART era,19 and the decline in mortality was universal among all demographic subgroups.20 The use of HAART itself was found to have an 80-91% reduction in AIDS, new opportunistic infections, or death in those with advanced disease.21 At the individual level, immune recovery has made it possible to treat many otherwise crippling opportunistic diseases, e.g. cryptosporidiosis, microsporidiosis, Kaposi's sarcoma, and progressive multifocal leukoencephalopathy (PML).

Monitoring of treatment (Algorithm 11(B))

Clinical evaluation

The patient should be seen frequently in the initial period of treatment. Adverse effects of medications are usually more severe and frequent in this stage. The physician should proactively manage such adverse effects. Diarrhoea with Kaletra might be controlled by a high fiber diet, Metamucil, calcium tablets or imodium. Patients should be warned to return to the clinic immediately when hypersensitivity reactions occur with NVP and ABC.

Immune reconstitution disease (see Chapter 15) is common with certain infections, such as mycobacteria, CMV (cytomegalovirus), Pneumocystis jiroveci (PCP) and viral hepatitis B and C. Presentations are typically atypical and require great clinical acumen for diagnosis and differentiation from treatment failure.

Viral load and CD4 count

Both viral load and CD4 count are important markers for monitoring, and are normally tested every three to 4 months. Notably the immediate response to treatment is assessed by viral load (VL). Effective treatment should decrease the viral load by at least 2 logs within 8 weeks and render the viral load undetectable within 6 months. Failure to achieve these milestones should prompt an immediate evaluation to exclude non-adherence, unfavourable drug interactions, drug malabsorption and superimposed infections or vaccination.

VL will remain undetectable as long as HAART is successful. Development of resistance is virtually nil in the absence of significant replication. However, 'viral blips' infrequently occur. These are intermittent episodes of detectable low-level viraemia, usually <1000/mL, which resolve spontaneously by returning to undetectable levels without a change of therapy. They are likely due to laboratory variation or viral release from latently infected cells. The challenge is to differentiate from early virolgoic failure and inadequate adherence. Careful and compassionate history taking, increasing the frequency of VL monitoring, and resistance testing for VL >1000/mL are necessary. The distinction is important as viral blip is not associated with viral resistance, whereas treatment failure usually involves resistance which becomes more extensive with time. Of note, intercurrent illness and vaccinations do not cause viral blips in patients on effective HAART, although the CD4 count may be temporarily depressed.

The current standard of care also demands pre- and post-treatment CD4 for monitoring. Pre-treatment CD4 indicates the timing of therapy, and post-treatment rise in CD4 assures an improved prognosis. It is noted that the lower the nadir CD4 count, the more limited is the eventual recovery and it is uncommon for it to return to pre-infection levels. Aside from clinical conditions, the CD4 count is the most important laboratory marker for initiating and stopping prophylactic therapy.

Biochemical and haematologic monitoring

The current antiretrovirals have significant toxicity. This would not be otherwise acceptable in the treatment of lesser disease. The range of antiretroviral toxicity is diverse and extensive as well as frequent monitoring is required. For example, the haematologic profile is monitored for toxicity of ZDV. Liver function tests are necessary for treatment with PI and NVP, and renal function tests with TDF, especially when given in combination with ddI and RTV. Cholesterol profile before and after treatment is also advisable. Most clinics undertake biochemical and haematologic monitoring for all patients on treatment, at a frequency no less than quarterly.

Lactic acidosis is a serious complication, but screening in asymptomatic patients is not recommended. However, in a patient with symptoms who are on nucleosides, a venous lactate and preferably arterial blood gas with lactate should be done immediately.

Resistance testing

It is not necessarily true that the viral population in the treatment-naïve consists solely of wild type. Patients who contracted the infection recently and in countries where antiretrovirals are available have a higher chance of harbouring primary resistance. Wild type virus is expected to overgrow resistant viruses because of a difference in replicative fitness. However, this process of revertant mutations may take a long time in the presence of extensive compensatory mutations that improve fitness.22

Consensus is forming that resistance testing should be performed in all patients as soon as possible after diagnosis to aid in the design of HAART. Resistance testing is also necessary when decrease in viral load is suboptimal following initiation of treatment, or viral load rebounds after initial suppression. The identification of mutations will guide the formulation of alternative therapy.

The methodology and applications of resistance testing is covered in Chapter 16.

Therapeutic drug monitoring

It is becoming obvious that therapeutic drug monitoring (TDM) could be useful in the monitoring of patients on ART. This is covered in Chapter 16.

Other considerations in the use of HAART

Interruption of therapy

Although STI as a strategy of treatment is inadvisable, HAART may need to be interrupted because the patient undergoes surgery, develops intercurrent illness or manifests significant toxicity. It is also not uncommon that a patient develops medication 'fatigue' and demands a drug holiday. All interruptions and subsequent re-start should be supervised by a specialist in HIV medicine, as there are risks involved. For example, after stopping an NNRTI-regimen, the prolonged half-lives of NVP and EFV will maintain their concentrations above the IC95 level for up to 3 weeks, greatly increasing the chance of resistance. Stopping TDF, 3TC, or FTC in a patient co-infected with hepatitis B may also cause a flare-up of hepatitis B. It has been reported that viral rebound upon discontinuation of therapy can result in the acute retroviral syndrome.

Tolerability and convenience

With current antiretrovirals, potency is generally not an issue for the wild type virus. It is toxicity and convenience, two major determinants of adherence, which have to be improved upon. Most recently introduced drugs have a low pill burden, and require only once or twice daily dosing by taking advantage of the pharmacokinetic enhancement by RTV. The availability of combination pills significantly reduces pill burden, one of which is a coformulation pill of TDF, FTC and EFV (Atripla®) which needs to be taken only once a day. However, it must be emphasised that there are considerable individual variations in tolerability and acceptance. Although there are 'recommended' regimens, individualised treatment takes precedence over a 'standardised' protocol.

Cost

Although it is cost effective, HIV infection has become the costliest disease to treat. Drugs alone average between HK$8,000 to 12,000 per month for the lifetime of a patient. Few societies will be able to sustain this cost if the rate of new infections remains unchanged. It will be prevention, not treatment, which constitutes the most effective response to the epidemic.

Algorithm 11(A)

Algorithm 11(B)

References

  1. Scientific Committee on AIDS. Recommended principles of antiretroviral therapy in HIV disease. Jan 2005. Available from http://www.aids.gov.hk (accessed June 26, 2006).

  2. Zhang H, Dornadula G, Beumont M, et al. Human immunodeficiency virus type 1 in the semen of men receiving highly active antiretroviral therapy. N Engl J Med 1998;339:1803-9.

  3. Wood E, Hogg RS, Yip B, Harrigan PR, O'Shaughnessy MV, Montaner JS. Effect of medication adherence on survival of HIV-infected adults who start highly active antiretroviral therapy when the CD4+ cell count is 0.200 to 0.350 x 10(9) cells/L. Ann Intern Med 2003;139:810-6.

  4. Dean GL, Edwards SG, Ives NJ, et al. Treatment of tuberculosis in HIV-infected persons in the era of highly active antiretroviral therapy. AIDS 2002;16:75-83.

  5. Kam KM, Leung WL, Kwok MY, Hung MY, Lee SS, Mak WP. Lymphocyte subpopulation reference ranges for monitoring human immunodeficiency virus-infected Chinese adults. Clin Diagn Lab Immunol 1996;3:326-30.

  6. Kam KM, Wong KH, Li PC, Lee SS, Leung WL, Kwok MY. Proposed CD4(+) T-cell criteria for staging human immunodeficiency virus-infected Chinese adults. Clin Immunol Immunopathol 1998;89:11-22.

  7. Wong KH, Chan KCW, Cheng KLS, Chan WK, Kam KM, Lee SS. HAART initiation in Chinese HIV/AIDS patients when their CD4 level is 100-150/ul is not associated with faster disease progression [Abstract H-575]. 44th ICAAC Washington, DC Oct 30-Nov 2, 2004

  8. DHHS. Guidelines for the use of antiretroviral agents in HIV-1 infected adults and adolescents. May 4, 2006. Available from http://www.aidsinfo.nih.gov (accessed June 23, 2006).

  9. Pomerantz RJ. Primary HIV-1 resistance: a new phase in the epidemic? JAMA 1999;282:1177-9.

  10. van Praag RM, Weverling GJ, Portegies P, et al. Enhanced penetration of indinavir in cerebrospinal fluid and semen after the addition of low-dose ritonavir. AIDS 2000;14:1187-94.

  11. Montaner JS, Reiss P, Cooper D, et al. A randomized, double-blind trial comparing combinations of nevirapine, didanosine, and zidovudine for HIV-infected patients: the INCAS Trial. Italy, The Netherlands, Canada and Australia Study. JAMA 1998;279:930-7.

  12. Staszewski S, Morales-Ramirez J, Tashima KT, et al. Efavirenz plus zidovudine and lamivudine, efavirenz plus indinavir, and indinavir plus zidovudine and lamivudine in the treatment of HIV-1 infection in adults. Study 006 Team. N Engl J Med 1999;341:1865-73.

  13. van Leth F, Phanuphak P, Ruxrungtham K, et al. Comparison of first-line antiretroviral therapy with regimens including nevirapine, efavirenz, or both drugs, plus stavudine and lamivudine: a randomised open-label trial, the 2NN Study. Lancet 2004;363:1253-63.

  14. DeJesus E, Herrera G, Teofilo E, et al. Abacavir versus zidovudine combined with lamivudine and efavirenz, for the treatment of antiretroviral-naive HIV-infected adults. Clin Infect Dis 2004;39:1038-46.

  15. Gallant JE, DeJesus E, Arribas JR, et al. Tenofovir DF, emtricitabine, and efavirenz vs. zidovudine, lamivudine, and efavirenz for HIV. N Engl J Med 2006;354:251-60.

  16. Negredo E, Molto J, Burger D, et al. Lopinavir/ritonavir plus nevirapine as a nucleoside-sparing approach in antiretroviral-experienced patients (NEKA study). J Acquir Immune Defic Syndr 2005;38:47-52.

  17. Duvivier C, Ghosn J, Assoumou L, et al. Lower rate of virological suppression in naive patients initiating HAART with NRTI-sparing regimen compared to standard NRTI-containing regimen: results from Hippocampe - ANRS 121 Trial. [Abstract PS1/3]. 10th EACS Nov 17, 2005 Dublin.

  18. Arribas JR, Pulido F, Delgado R, et al. Lopinavir/ritonavir as single-drug therapy for maintenance of HIV-1 viral suppression: 48-week results of a randomized, controlled, open-label, proof-of-concept pilot clinical trial (OK Study). J Acquir Immune Defic Syndr 2005;40:280-7.

  19. Wong KH, Chan KC, Lee SS. Delayed progression to death and to AIDS in a Hong Kong cohort of patients with advanced HIV type 1 disease during the era of highly active antiretroviral therapy. Clin Infect Dis 2004;39:853-60.

  20. Chan KC, Wong KH, Lee SS. Universal decline in mortality in patients with advanced HIV-1 disease in various demographic subpopulations after the introduction of HAART in Hong Kong, from 1993 to 2002. HIV Med 2006;7:186-92.

  21. Chan CW, Cheng LS, Chan WK, Wong KH. Highly active antiretroviral therapy per se decreased mortality and morbidity of advanced human immunodeficiency virus disease in Hong Kong. Chin Med J (Engl) 2005;118:1338-45.

  22. Chan KC, Galli RA, Montaner JS, Harrigan PR. Prolonged retention of drug resistance mutations and rapid disease progression in the absence of therapy after primary HIV infection. AIDS 2003;17:1256-8.