Clinical monitoring

Detecting primary infection

In the absence of laboratory capabilities, several regions have developed algorithms for determining acute HIV infection (AHI). Identifying acute infection has significant implications for curbing the spread of HIV infection, particularly in resource-limited settings.

US and African researchers have developed an algorithm, based on rapid test results, symptoms, and risk behaviours, which makes it possible to accurately detect acute HIV infection without widespread use of HIV RNA assays. In sub-Saharan Africa, roughly 1500 people were screened for AHI using two rapid HIV antibody tests. Those who had discordant results or two positive results received HIV RNA or p24 antigen tests.

Acute HIV infection was strongly associated with discordant rapid HIV tests and genital ulcer disease (GUD). Predictor indications in this population were weighted one point for fever, body ache, and more than one sexual partner in two months; two for diarrhoea lasting at least a month and GUD; and four for discordant rapid tests. A resultant risk score of two or greater identified 95% of actual cases of AHI (61% specificity) while requiring the use of resource-intensive HIV RNA or p24 antigen tests on only 41% of individuals presenting for testing.¹

Risk score algorithms need to be developed and evaluated according to setting and population - factoring in subtype, symptoms, risk behaviours, assays used, and other meaningful variables in that area. Following validation, the algorithm could provide rapid, reliable, and cost-effective AHI detection.

Baseline examination

Valuable information can be gained, even without the benefit of laboratory testing, by taking a good medical history and performing a physical exam. The World Health Organization guide to disease staging can be very useful in resource-limited settings.

It is important to ask about any current symptoms and to record information on, among other items, weight loss; diarrhoea; bacterial infections, including TB; ongoing fevers; and sexually transmitted infections. Recording a person's weight when they are first seen, and at intervals after that, is a simple example of monitoring that can identify serious illness or record progress on treatment.

A good physical exam should include vital signs and weight; HEENT (head, eyes, ears, nose, and throat), including checking for the presence of thrush, hairy leukoplakia, and retinal lesions; and checking for the presence of lymphadenopathy; signs of hepatomegaly or splenomegaly; peripheral oedema; peripheral neuropathy and focal neurologic abnormalities; evidence of pulmonary problems; and evidence of skin problems, including past or current herpes zoster.

It is important to make sure that there is no undetected serious illness, in addition to HIV, which could be treated or that might affect antiretroviral choices. Tuberculosis screening is vital to identify people who can benefit from curative or preventive treatment for TB.

Anaemia (low haemoglobin count) has been identified in a number of populations as a risk factor for mortality in HIV disease. It may point to nutritional deficiencies, the presence of treatable parasitic diseases, or a side-effect of antiretroviral therapy (ART).

Ongoing monitoring

Patients not on HIV treatment should be seen in clinic every three to six months. If symptoms develop at any time, the clinician needs to follow up immediately and try to ascertain if the problem is HIV-related. CD4 cell counts should be performed every six months, when feasible.

Patients on ART should be monitored for:

Clinical improvement (resolution of fever, diarrhoea, weight gain etc.)
Potential clinical disease progression (fevers, diarrhoea, weight loss, occurrence of opportunistic infection and/or malignancy)
Potential drug toxicity and/or treatment side-effects
Drug adherence
For women, occurrence of cervical dysplasia, menstrual disorders, or pregnancy.

Expanded access to antiretroviral treatment should be offered in all settings, regardless of the availability of appropriate laboratory facilities, according to the results of several recent studies.

Comparing approaches to monitoring

The main problem with existing and approved viral load assays is making them affordable and available in resource-limited areas. Barriers are the cost of the technology, lack of laboratory infrastructure, and availability of skilled technicians.

Options being explored are the use of a p24 assay, as this has been shown to be a reliable prognostic indicator of disease progression. Currently, the PerkinElmer Ultrasensitive p24 is available. This assay is fast and easy, but still under evaluation for clinical utility.

The ExaVir Load Version 3 mentioned above has been used in several studies and found to be as accurate as the other commercial assays that measure RNA using a PCR, bDNA, or NASBA platform. The Homebrew Real-Time PCR and Primagen Retina Rainbow/NucliSens EasyQ are two other assays being evaluated for use in low-resource areas.

A review of HIV-1 viral load assays available in resource-limited areas has been done by the Forum for Collaborative HIV Research and is available at PLoS (Public Library of Science) online.²

In resource-limited areas, reports from research- and community-based settings have reported on the efficacy, or lack thereof, of the following approaches to monitoring viral suppression. Summarised below are just a few of these reports. The research on use of total lymphocyte count was discussed earlier in the section Other immunological markers.

Using HIV symptoms

A mathematical modelling program that used survival as an outcome measure, rather than treatment failure, found that the use of HIV symptoms to monitor patients was almost as effective as the use of CD4 cells counts or viral load. Similar outcomes were found when antiretroviral therapy was changed for any of the following reasons: viral load greater than 500 copies/ml; a 50% decrease in CD4 cell count from peak; or when two new WHO stage 3 events or one WHO stage 4 event occurred.³

If virological failure (defined here as a viral load above 1000 copies/ml) was the outcome measure, then 87% of cases would have been detected at year one by viral load, 32% using two WHO stage 3 or one stage 4 event, and 25% using a CD4 cell decline of 33% in six months. If a viral load above 500 copies/ml defined virological failure, then only 16% of cases would have been detected at one year.

Clinical monitoring alone

Modifying this view are the recent results of the DART study, in which over 3000 patients initiating ART were followed for nearly five years. Participants were randomised either to receive treatment with routine laboratory monitoring, or to receive treatment with clinical monitoring supported by laboratory tests only where the clinician needed more information about a patient’s condition in order to make a clinical decision, or where a grade 4 toxicity was detected.

Differences in disease progression became apparent only after the second year of treatment was completed, leading investigators to suggest that clinical monitoring alone was a reasonable approach during the first two years of ART. After that time, they recommended using clinical criteria and CD4 counts to guide treatment changes.⁴ [ref] [ref]

Adherence

In a South African, multi-country observational cohort of nearly 2000 patients, measuring adherence turned out to be as accurate as CD4 cell counts for detecting virologic failure.

The manner of assessing adherence was simple: the number of pharmacy prescriptions filled was divided by the number of months on therapy over a defined period. In the first year after initiating antiretroviral treatment, CD4 cell count and adherence were equally accurate at detecting breakthrough viraemia. Adherence level more accurately predicted virologic failure than did CD4 count change.⁵

Since adherence lapses precede virological failure, this strategy identifies patients at high risk of treatment failure and allows the time for a possible intervention before viral load increases lead to drug resistance.

Other studies have suggested the utility of behavioural over biologic monitoring to in resource-limited areas. Continuous adherence monitoring may be a cost-effective way in which to forestall viral rebound.⁶ ⁷

The Swiss HIV Cohort study reported that there is a significant association between adherence and optimal viral suppression. Adherence was gauged by answers to two questions: one about the frequency of missed doses in the prior four weeks and the other asking whether there had been any 24-hour period in the past month in which no medication at all was taken.⁸

CD4 cell count

In the absence of viral load testing, World Health Organization (WHO) guidelines suggest that a CD4 count decline of 50% from the previous peak, or a one-third decline in the previous six months, should be considered a trigger for changing treatment. However, there is limited clinical evidence in resource-limited settings that CD4 cell counts are accurate indicators of viral suppression on treatment.

To address this question, researchers in the Ugandan Home-Based AIDS Care Project analysed the relationship between immunologic and virologic markers amongst 1000 patients initiating treatment.

They found that the proportion with no increase in CD4 count from baseline was the same between those with suppressed or unsuppressed viral load at 6, 18, and 24 months after ART initiation.

CD4 cell count alone had a sensitivity of 8% in this analysis, so it would have missed 92% of those patients who were not suppressing their viral load. With an 11% positive predictive value, 89 out of 100 people identified as failing treatment would have had actually had suppressed virus and their regimens would have been unnecessarily changed as a result of the CD4 count results.⁹

Frequency of viral load testing

Viral load testing once or twice a year had the same impact on patient health outcomes as more frequent testing in a study of over 2300 patients from sites across the Asia-Pacific region. However, a 35% increased risk of severe symptomatic HIV disease and death was reported in sites with less than once-yearly viral load testing.

Data were drawn from 2333 patients in the TREAT Asia Observational Database (TAHOD). Less than annual viral load testing was associated with reduced odds of viral suppression and significantly higher rates of disease progression. The authors believe this reflects lack of capacity at individual sites to identify for viral load testing those at high risk of disease progression, a challenge in resource-poor settings. However, there was no significant difference in disease progression between once or twice a year viral load testing and testing three times or more a year.¹⁰

Viral load testing before switching regimens

A year-long South African study found that WHO clinical and CD4 criteria had poor sensitivity and specificity in determining treatment failure in over 300 persons initiating ART.

The sensitivity of CD4 criteria in detecting treatment failure was 21% and specificity was 96%. Clinical criteria had a sensitivity of 15% and specificity of 88%. The use of clinical monitoring and CD4 counts together resulted in very low predictive values.

The majority of patients who were identified as having treatment failure over the course of twelve months actually had adequate viral suppression. These results have ongoing consequences considering limited ARV availability and the cost of second-line treatment.

In this case, the authors suggest that it would be less expensive to do a viral load test than to switch treatment regimens unnecessarily. The same study also found that a 10% weight loss and the occurrence of TB were not indicative of treatment failure.¹¹

A small Chinese study also found limited utility of CD4 count and clinical evaluation in determining treatment failure.¹²

Viral load testing as indicated by adherence and CD4 counts

A clinical algorithm using targeted viral load testing was constructed to identify patients for second-line ART. This algorithm was compared with the World Health Organization (WHO) guidelines, which use clinical and immunological criteria to identify failure in the absence of viral load testing.

Combining adherence failure (interruption >2 days) and CD4 failure (30% fall from peak) had a sensitivity of 67% for a viral load of >1000 copies/ml and a specificity of 82%; identifying 22% of patients for viral load testing. The WHO-based model had a sensitivity of 31%, specificity of 87%, and would result in 14% of those with viral suppression (<400 copies/ml) being switched to second-line ART.¹³

Other laboratory monitoring

Another study examined the clinical benefit and cost-effectiveness of routine monitoring for asymptomatic laboratory abnormalities, versus testing only those presenting with symptoms of toxicity, among patients on antiretroviral treatment in Haiti.¹⁴

Routine laboratory monitoring in asymptomatic patients was only clinically beneficial and cost-effective in patients with asymptomatic anaemia, and in patients concurrently taking tuberculosis medication because they were at increased risk for liver abnormalities. The investigators noted that findings could vary with factors such as antiretroviral agents in use and co-incident health conditions.

References

Powers KA et al. Improved detection of acute HIV-1 infection in sub-Saharan Africa: development of a risk score algorithm. AIDS 21(16): 2237-2242, 2007
Fiscus SA et al. HIV-1 viral load assays for resource-limited settings. PLoS Med 3(10): e417. doi: 10.1371/journal.pmed.0030417, 2006
Phillips AN et al. Outcomes from monitoring of patients on antiretroviral therapy in resource-limited settings with viral load, CD4 cell count, or clinical observation alone: a computer simulation model. Lancet 371: 1443-1451, 2008
Mugyenyi P et al. Impact of routine laboratory monitoring over 5 years after antiretroviral therapy (ART) initiation on clinical disease progression of HIV-infected African adults: the DART Trial final results. Fifth International AIDS Society Conference on HIV Pathogenesis, Treatment and Prevention, Cape Town, abstract TuSS103, 2009
Bisson GP et al. Pharmacy refill adherence compared with CD4 count changes for monitoring HIV-infected adults on antiretroviral therapy. PLoS Med (5): e109 doi:10.1371/journal.pmed.0050109, 2008
Bangsberg DR A paradigm shift to prevent HIV drug resistance. PLoS Med 5(5): e111, 2008
Mannheimer S et al. The consistency of adherence to antiretroviral therapy predicts biologic outcomes for human immunodeficiency virus-infected persons in clinical trials. Clin Inf Dis 34: 1115-21, 2002
Glass TR et al. Correlates of self-reported nonadherence to antiretroviral therapy in HIV-infected patients: the Swiss HIV Cohort Study. J Acquir Immune Defic Syndr 41(3): 385-392, 2006
Moore DM et al. CD4+ T-cell count monitoring does not accurately identify HIV-infected adults with virologic failure receiving antiretroviral therapy. J Acquir Immune Defic Syndr 49: 477-484, 2008
Oyomopito R et al. Measures of site resourcing predict virologic suppression, immunologic response and HIV disease progression following highly active antiretroviral therapy (HAART) in the TREAT Asia observational database (TAHOD). HIV Medicine 11(8): 519-529, 2010
Mee P et al. Evaluation of the WHO criteria for antiretroviral treatment failure among adults in South Africa. AIDS: 22(15): 1971-1977, 2008
Chaiwarith R et al. Sensitivity and specificity of using CD4+ measurement and clinical evaluation to determine antiretroviral treatment failure in Thailand. Int J Infect Dis 11:413-416, 2007
Meya D et al. Development and evaluation of a clinical algorithm to monitor patients on antiretrovirals in resource-limited settings using adherence, clinical and CD4 cell count criteria. J Int AIDS Soc 12(1): 3, 2009
Koenig SR et al. Clinical impact and cost of monitoring for asymptomatic laboratory abnormalities among patients receiving antiretroviral therapy in a resource-poor setting. Clin Infect Dis advance online edition July, 2010