University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Evaluation of the WHO criteria for antiretroviral treatment failure among adults in South Africa
Global Health Sciences Literature Digest
Published April 17, 2009
Journal Article

Mee P, Fielding KL, Charalambous S, Churchyard GJ, Grant AD. Evaluation of the WHO criteria for antiretroviral treatment failure among adults in South Africa. AIDS 2008;22:1971-7.

In Context

Persons receiving antiretroviral therapy (ART) frequently develop treatment resistance. A key determinant of treatment failure is increases in viral load, tests for which generally are not available in developing countries. As such, the WHO developed guidelines for ARV treatment failure that are based upon CD4 counts and clinical criteria, as these measures are generally available;(1) however, objective evaluation of these criteria has not been undertaken.


To evaluate the WHO criteria for ART failure


A work-based ARV treatment program in South Africa

Study Design

Prospective cohort


Antiretroviral-naïve patients who initiated ART between November 13, 2002 and August 7, 2004 and who had a follow-up visit that included viral load testing 12 months (range 300-450 days) after beginning ARV therapy, and whose viral load result was available for review. Patients who stopped or switched treatment during that time were excluded. Patients who temporarily discontinued treatment but resumed with the first-line therapy were included.


The sensitivity, specificity and positive and negative predictive values of the WHO guidelines for treatment failure using definitions based upon viral load measurements as the gold standard


Treatment failure using viral load standards was defined as a) definite if a single HIV viral load greater than 10,000 copies/mL at the 12-month visit; b) probable if a single viral load at 12 months was greater than 1,000 copies/mL or if a viral load at 12 months greater than 400 copies/mL and a repeat viral load measurement 30 days later also greater than 400 copies/mL.

Treatment failure as defined by the WHO standards was CD4 count at 12 months that was below the pretreatment baseline value (defined in this study as the CD4 count closest to the start of treatment and within 90 days of initiating treatment) or a fall in CD4 count to less than 50% of the maximum CD4 cell count while on therapy. A CD4 count of less than 100 cells/µL was a definite failure and a count greater than 200 cells/µL was never a failure.

In a secondary analysis, treatment failure was based upon CD4 counts at the six-month visit and used the same CD4 values for treatment failure at 12 months.

Treatment failure based upon clinical criteria was defined as either new or recurrent WHO-defined stage 3 or stage 4 conditions at six months or later after the start of treatment. Severe weight loss was defined as a loss of at least 10% of the maximum weight after starting ART. A severe bacterial infection was one that resulted in hospitalization. The medical records of all patients meeting clinical criteria for treatment failure were reviewed by an experienced HIV clinician to identify any conditions that might have occurred for reasons other than treatment failure. Clinical conditions occurring within the first six months of treatment were not considered failures.


There were 676 ART-naïve patients with baseline CD4 data. Of these, 423 (63%) remained in the program and on ART throughout the 12 months. Two patients changed to second-line therapy and were excluded. There were 97 people who did not have a 12-month viral load or CD4 test result. The remaining 324 patients were primarily male, had a median age of 40.2 years, weighed 65 kg, had a viral load of 47,503 copies/mL, and had a CD4 count of 154 cells/µL. Of these patients, five had only a single viral load between 4,001 and 1,000 copies/mL and were excluded.

Twelve months after starting treatment, 19 (6%) patients met the criteria for definite virologic treatment failure and 14 (4.4%) met the criteria for probable failure. There were 19 (6%) who met the CD4 criteria for treatment failure, 12 (3.8%) had a CD4 cell count at 12 months of less than the value before starting ART, six (1.9%) had a CD4 cell count of less than 50% of the maximum during therapy, and nine (2.8%) had a CD4 cell count of less than 100 cells/mL. Forty (12.5%) patients met the clinical criteria for failure; 18 (5.6%) from weight loss, 11 (3.4%) from TB, and the remainder from other stage 3 or 4 conditions.

CD4 criteria had sensitivity of 21.2% and a specificity of 95.8% in detecting virologic failure, and clinical criteria had sensitivity of 15.2% and specificity of 88.1%. The positive predictive value of CD4 and clinical criteria in detecting virologic failure were 36.8 and 12.8%, respectively.


The WHO clinical and CD4 criteria for ART failure have poor sensitivity, specificity, and predictive value. The low specificities and positive predictive values are likely to result in misclassification of patients who have adequate virologic response as treatment failures.

Quality rating

This is a high-quality study. The cohort is representative of people treated in employee-based clinics, but probably not of persons treated in other settings. Ascertainment of exposure (treatment) and outcome (treatment failure) was very good. Some patients were excluded based upon missing data, and it is not clear if these people differed from those who remained in the study. Follow-up time was appropriate, and a sufficient proportion of persons completed the study.

Programmatic Implications

The findings from this study confirm others(2,3,4) and support calls for more accurate algorithms to detect treatment failure in developing countries.(5) These findings are very important as they demonstrate that the current recommendations for diagnosing ART failure are quite poor. At times, clinical recommendations for use in resource-constrained areas try to draw on available resources even when there is insufficient evidence of their usefulness. This study demonstrates the hazards of developing and applying such recommendations, as use of these criteria is likely to result in unnecessary changes to therapy. It is both cost-effective and medically preferable to maintain patients on a successful first-line antiretroviral regimen.


  1. World Health Organization. World Health Organization Antiretroviral therapy in adults and adolescents - recommendations for a public health approach - 2006 revision.
  2. Moore DM, Mermin J, Awor A, Yip B, Hogg RS, Montaner JS. Performance of immunologic responses in predicting viral load suppression: implications for monitoring patients in resource limited settings. J Acquir Immune Defic Syndr 2006;43:436-9.
  3. Bisson GP, Gross R, Strom JB, et al. Diagnostic accuracy of CD4 cell count increase for virologic response after initiating highly active antiretroviral therapy. AIDS 2006;20:1613-9.
  4. Chaiwarith R, Wachirakaphan C, Kotarathititum W, Praparatanaphan J, Sirisanthana T, Supparatpinyo K. Sensitivity and specificity of using CD4+ measurement and clinical evaluation to determine antiretroviral treatment failure in Thailand. Int J Infect Dis 2007;11:413-6.
  5. Colebunders R, Moses KR, Laurence J, et al. A new model to monitor the virological efficacy of antiretroviral treatment in resource-poor countries. Lancet Infect Dis 2006;6:53-9.