This was a cross-sectional study of 496 Ugandans established on NNRTI-based ART. We evaluated combinations of adherence, clinical and laboratory variables to determine viral failure.

This study was performed at the adult clinic of the Infectious Disease Institute (IDI), Mulago Hospital, Makerere University in Kampala, Uganda. The IDI is one of Uganda's largest HIV treatment centres with more than 10,000 active patients and more than 5000 patients currently on free ART 18. The IDI is supported by a College of American Pathologists-certified laboratory and is able to perform CD4 cell counts and viral load testing on site.

Patients were screened and included in the study if they were HIV-1 positive, aged >18 years, established on first-line NNRTI-based ART for ≥ six months and did not have viral loads monitored as per routine clinic practice. Patients with acute illness were excluded from the study.

From February 2006 to June 2006, 500 patients were enrolled at a rate of approximately 10 patients per clinic day. Patients were randomly selected from the clinic reception using a list of random numbers.

The study doctor carried out a structured interview and chart review using a study questionnaire. The questionnaire included detailed questions about treatment history, adherence to ART, clinical events and changes in laboratory parameters, including CD4 cell count since the start of treatment. CD4 cell counts are routinely ordered at the IDI every six months, with additional measurements taken if judged necessary by the treating physician.

Adherence was measured by self report, using a modified Adult AIDS Clinical Trials Group adherence questionnaire validated in our setting 1920. Participants were asked to report adherence patterns in the three days prior to enrolment, four weeks prior to enrolment, and since the initiation of ART. A visual analogue scale, as well as a question on whether treatment had ever been interrupted for more than two days, was included to assess adherence in the four weeks prior to enrolment and since the initiation of ART 21.

A blood sample was then taken for a complete blood count (ACT diff2 – Beckman Coulter, California, USA), CD4cell count and percentage (FACScalibur – Becton Dickenson, New Jersey, USA), and viral load (Amplicor HIV-1 Monitor v1.5 – Roche, Switzerland). The lower limit of detection for viral load was 400 copies/mL. An additional plasma sample was stored for each patient. Participants found to have a viral load of >1000 copies/mL underwent a genotypic resistance test (Trugene HIV-1 Genotyping Kit, Visible Genetics – Bayer Diagnostics, Leverkusen, Germany).

Examined variables included: months on ART; history of antiretroviral regimen limited to dual or triple nucleoside reverse transcriptase inhibitor therapy; history of maternal single-dose nevirapine to prevent vertical transmission; history of ever paying for ART; missing any ART during the last 30 days of treatment; ever missing ART for more than two days, current weight less than baseline weight; HIV-related symptoms, including prurigo and onset or relapse of opportunistic infection (OI); CD4 cell count change from baseline; 30% fall in CD4 cell count from on-treatment peak value; and WHO immunologic failure criteria (fall of CD4 count to pre-therapy baseline or below, 50% fall from on-treatment peak value, and persistent CD4 cell count of <100 cells/mm³).

A new or recurrent OI was defined according to 2006 WHO guidelines 17 as a WHO Stage 4 event (plus any severe bacterial infections or pulmonary tuberculosis) occurring six months after initiation of ART.

We used χ² and Fisher's exact tests to compare categorical data, and the Kruskal-Wallis test to compare continuous variables. P values of < 0.05 were considered statistically significant. Univariate and multivariate logistic regression analysis was used to model variables associated with viral failure (>400 copies/mL).

We constructed the multivariate model by entering variables that were significant in the univariate analysis. To address multicollinearity, we examined variables that were strongly correlated and chose the variable with the greatest magnitude of association with viral failure to include in the multivariate model.

Variables in the final model were: gender; age; months on ART; history of paying for ART; ever missed more than two days of ART; 30% fall from peak CD4 cell count; and new or recurrent OI. A monitoring algorithm was then constructed using those parameters significantly associated with viral failure by multivariate logistic regression analysis.

Finally, we compared the ability of the regression-based algorithm and an algorithm using the WHO clinical and immunological treatment failure criteria 17 to classify patients according to viral status. We calculated sensitivity, specificity, positive and negative predictive value to determine viral failure <1000 copies/mL. Data were analysed using SAS version 8.2 (Cary, NC, USA).

Informed consent was obtained from all the participants. Ethical approval for this study was obtained from the National Council of Science and Technology (Uganda) and from the National Institute of Allergy & Infectious Diseases (USA).

Five hundred participants were enrolled, of which 496 had completed questionnaires and viral load results. Median age was 38.4 years (IQR, 33.5 to 43.7 years), and 311 (62.7%) were women. Forty-nine (9.9%) patients had a detectable viral load (>400 copies/mL). Thirty-nine (7.9%) patients had a viral load of >1000 copies/mL. Detectable viral loads ranged from 416 to 447,000 copies/mL.

The median duration of ART was 13 months (IQR, 10 to 16 months). The median CD4 cell count at baseline, before starting ART, was 90 cells/mm³ (IQR 35 to 156 cells/mm³). The median CD4 cell count gain on treatment was 138 cells/mm³ (IQR, 76 to 224 cells/mm³).

Eleven participants developed a new or recurrent OI on ART. These included Pneumocystis jiroveci pneumonia (N = 2), cryptococcal meningitis (N = 3), pulmonary tuberculosis (N = 3), extrapulmonary tuberculosis (N = 1), Kaposi's sarcoma (N = 1), and severe bacterial infection (N = 2). One participant suffered episodes of both Pneumocystis jiroveci pneumonia and pulmonary tuberculosis. Of these 11, only three had viral failure, including two participants with cryptococcal meningitis and one participant with severe bacterial infection.

Table 1 summarizes the univariate results for adherence patterns, clinical events and laboratory variables associated with viral failure. Odds ratio for self report of ART missed in the last 30 days was 1.9 (95% CI 0.9 to 4.1) and for ever missed more than two days of ART was 6.3 (95% CI 3.4 to 11.8).

CD4 cell count was measured by change in CD4 cell count from baseline, 30% fall and 50% fall from maximum achieved, persistent CD4 cell count of <100 cells/mm3, and CD4 cell count at study visit below baseline. WHO criteria were evaluated by univariate analysis of immunologic (CD4 cell count-based) criteria (OR, 2.1; 95% CI 1.0 to 4.3) and immunologic criteria and Stage 4 disease (OR, 2.1; 95% 1.0 to 4.2).

In the multivariate logistic regression model, ever missing ART for more than two days (OR, 5.2; 95% CI, 2.5 to 11.0) and 30% fall from peak CD4 cell count (OR, 3.9; 95% CI, 1.6 to 9.4) were significantly associated with viral failure (>400 copies/mL) after adjustment for gender, age, months on ART, history of paying for ART, and new or recurrent OI. Due to missing data for months on ART (N = 492) and 30% fall from peak CD4 cell count (N = 495), the multivariate results are based on 491 participants.

The parameters significantly associated with viral failure (>1000 copies/mL) by multivariate logistic regression, ever missing ART for more than two days, and 30% fall in CD4 cell count were used to construct a monitoring algorithm (Figure 1a). Participants who met either criteria were classified as "failure possible" and were recommended for viral load testing (N = 112).

Those patients without either of these parameters were classified as "failure unlikely" and were not recommended for viral load testing (N = 384). According to the regression-based algorithm, no combination of parameters was predictive of viral failure. Therefore it was not possible to categorize participants as "failure likely" or recommend a second-line ART regimen without viral load testing.

In a WHO-based algorithm (Figure 1b), patients with immunologic failure and Stage 4 disease (not including lymph node TB, uncomplicated TB pleural disease, oesophageal candidiasis, and recurrent bacterial pneumonia occurring after six months of therapy) were recommended to switch to second-line ART without viral load testing (N = 74). Patients without these criteria were recommended to continue first-line ART (N = 422).

The performance of the algorithms was assessed by sensitivity, specificity, and positive and negative predictive value, and then compared to an algorithm based on the WHO treatment failure criteria (Table 2). The regression-based algorithm identified patients with viral failure >1000 copies/mL with sensitivity of 67% and specificity of 82%, and identified 22% of patients for viral load testing. Thirty-three percent of patients with viral failure would continue first-line ART.

Sensitivity of the WHO-based algorithm was 31% and specificity was 87%. This approach would not involve any viral load testing, would leave 69% of patients with viral failure on first-line ART, and would inappropriately switch 14% of those with viral suppression (<400 copies/mL) to second-line ART.

In a modification to the WHO-based algorithm, those patients meeting the criteria of CD4 failure or Stage 4 disease could have viral load testing rather than switching to second-line ART. Results were similar if viral load of either >400 or >10,000 copies/mL determined viral failure (data not shown).

Drug resistance testing was performed in 38 of the 39 participants with viral load of >1000 copies/mL (Table 3). In four participants, no mutations were identified, and in three participants, it was not possible to amplify the virus.

Significant mutations of the reverse transcriptase region were identified in 31 of the 35 (89%) participant samples in which amplification was successful 22. All but one of these participants had mutations conferring resistance to either lamivudine (M184V) and/or NNRTIs (K103N, V108I, Y181C, G190A, P225H). Twelve participants (34%) had one or more thymidine analogue mutations (TAMs).