Background:   The South African government ART roll-out program, initiated in April 2004, has enrolled >150,000 patients to date. This is the first study describing pol mutation patterns arising in patients failing either the first or second-line ART regimens from 2 Johannesburg clinics. Both clinics report an overall therapy failure rate of approximately 2% among the 10,000 patients treated to date.

Methods:  Clinic 1 routinely sends samples for genotyping upon virological failure whereas clinic 2 only sends problem clinical cases due to resource constraints. Plasma from 100 patients (74 clinic 1; 26 clinic 2) was used to extract HIV RNA, protease (PR) and reverse transcriptase (RT) regions were sequenced, processed by the ViroScore database and analyzed for mutation frequencies. Subtype was designated using the Rega HIV-1 subtyping tool.

Results:  Patients were on HAART regimens containing stavudine (d4T) + lamivudine (3TC) + efavirenz (EFV) or nevirapine (NVP) (85%), or zidovudine (AZT) + didanosine (ddI) + Kaletra (15%). Patients attending clinic 1 and clinic 2 had average CD4 counts of 198 and 89 cells/µL, and viral loads of 29,404 and 106,854 copies/mL, respectively. Of all patients, 68% harbored known ART drug-resistance mutations. Major RT mutations from clinic 1 were:  D67N (14%), M184V (43%), K103N (28%), V106M (17%), G190A (18%). For clinic 2 they were:  M41L, K65R and T215Y (12%); D67N (27%); K70R (23%); M184V (46%); K219Q (15%); K101E and G190A (12%); K103N (42%); V106M (19%); and P225H (19%). Interestingly, 22% of clinic 1 patients and 35% of clinic 2 patients had nucleoside analog mutations; 11 patients harbored primary protease inhibitors (PI) mutations (L10I, M46I, G48V, I54V, A71V, V82A, and L90M), 3 of whom were not on PI-containing regimens. A high number of secondary PR mutations were observed among all patients. Subtype analysis revealed that 97% of samples were subtype C.

Conclusions:  Preliminary HIV-1 subtype C ART drug-resistance data emerging from 2 clinics in Johannesburg show mutation patterns similar to subtype B. Several important differences were noted, such as the presence of V106M, and higher prevalence rates of K65R (clinic 2), G190A and P225H (clinic 2). Thus, subtype C-specific resistance algorithms are urgently needed to avoid biased interpretation. The emergence of K65R mutations will limit the use of tenofovir (TDF) in second-line treatment. Treatment guidelines without viral load monitoring or defining treatment failure at viral load >5000copies/mL may limit second line treatment options with nucleoside analogs.