667
Emergence of Drug Resistance on First-line Combination Therapy in the Swiss HIV Cohort Study: A Comparison between Boosted PI and NNRTI
V von Wyl1, S Yerly2, J Böni3, J Schüpbach3, P Burgisser4, T Klimkait5, M Rickenbach6, L Perrin2, B Ledergerber1, Huldrych Günthard*1, and the Swiss HIV Cohort Study (SHCS)
1Univ Hosp, Zurich, Switzerland; 2Geneva Univ Hosp, Switzerland; 3Univ of Zurich, Switzerland; 4Lausanne Univ Hosp, Switzerland; 5Univ of Basel, Switzerland; and 6Swiss HIV Cohort Study Data Ctr, Lausanne
Background: Guidelines recommend initiating combination ART
(cART) with 2 nucleoside reverse transcriptase
inhibitors (NRTI) and either a protease inhibitor boosted with ritonivir (PI/r) or 1 non-NRTI (NNRTI). We investigated
failure rates and emergence of drug resistance for these 2 cART
types in previously untreated patients from the Swiss HIV Cohort Study (SHCS).
Methods: Patients were selected if they had initiated cART containing PI/r or NNRTI between January 1999 and December
2005 and experienced virological failure of ≥1
on-treatment viral load of ≥500 copies/mL after
≥180 days of continuous therapy. The International AIDS Society (IAS) -USA
mutation list of Fall 2006 and the Stanford algorithm
4.2.0 were used to assess resistance. For the analysis of class resistance
drugs were grouped into the following 4 classes: lamivudine/emtricitabine
(3TC/FTC); NRTI excluding 3TC/FTC; PI; NNRTI. Class resistance was defined as
resistance to ≥1 drug in a class.
Results: Of 711 patients on NNRTI 42 (5.9%) experienced virological failure, corresponding to a crude incidence
rate (95%CI) of 2.6 (1.9 to 3.5). Among 444 patients on PI/r, the number of virological failures was 22 (5.0%) and the crude incidence
was 2.9 (1.9 to 4.4). Resistance tests could be retrieved for 12 of 22 (54.5%)
failing patients of the PI/r group and 30 of 42 (71.4%) of the NNRTI group. Resistance
tests at baseline could be retrieved for 8 of 22 (36.4%) in the PI/r group and 18
of 42 (42.9%). Except for 1 reverse transcriptase (RT) 103N mutation in a
patient on PI/r, no baseline resistance was observed. Around the time of
failure, IAS mutations were found in 4 of 12 (33.3%) and 22 of 30 (73.3%,
Fishers exact p = 0.02) in the PI/r
group and the NNRTI group, respectively, affecting a mean, median [IQR] of
0.42, 0[0 to 1] classes in the PI/r group and 1.47, 2 (0 to 2) in the NNRTI
group (Mann-Whitney U, p = 0.0021). This
difference remained statistically significant when this analysis was repeated
using a Stanford genotypic sensitivity score (GSS) >15 or a GSS >59 as
cut-offs. Of patients on NNRTI 20 (67%) lost ≥2 classes upon first-line
failure as opposed to only 1 (8.3%) in the PI/r group. The most frequent
mutation was RT M184V, observed in 20 of 30 (66.6%) patients on NNRTI and 3 of 12
(25%) on PI/r (Fishers exact p = 0.017).
Conclusions: While PI/r containing cART
has similar potency like NNRTI regimens, they lead to less resistance in case
of virological failure. These data demonstrate that
in addition to drug potency, genetic barrier should also be considered when
choosing first-line treatment.
|
