Boston, Massachusetts
March 4–7, 2018


Conference Dates and Location: 
February 13–16, 2017 | Seattle, Washington
Abstract Number: 



Josep M. Llibre1, Anna Esteve2, Jose M. Miro3, Gracia Mateo4, Adrià Curran5, Daniel Podzamczer6, Melchor Riera7, Francesc Homar8, Luis Force9

1Univ Hosp Germans Trias, Barcelona, Spain,2Cntr d'Estudis Epidemiològics Sobre les ITS i Sida de Catalunya, Barcelona, Spain,3Univ of Barcelona, Barcelona, Spain,4Hosp Sant Pau, Barcelona, Spain,5Hosp Vall d'Hebró, Barcelona, Spain,6Hosp Univ de Bellvitge, Barcelona, Spain,7Hosp Univ Son Espases, Palma de Mallorca, Spain,8Hosp Son Llàtzer, Palma de Mallorca, Spain,9Hosp de Mataró, Mataró, Spain

Abstract Body: 

The rates of discontinuation (D/C) due to adverse events (AEs) of the integrase strand transfer inhibitors (INSTI) dolutegravir (DTG), raltegravir (RAL) and cobicistat-boosted elvitegravir (EVG/c) have been very low in randomized clinical trials. However, some real-life retrospective series have reported unexpectedly high rates of D/C due to AEs, particularly with DTG. We aimed to compare the D/C rates due to AEs of the three INSTI inhibitors in a prospective multicenter cohort.

The PISCIS Cohort is an ongoing observational study that includes about 21000 HIV-infected patients aged ≥16 years from 10 hospitals in Catalonia and 2 in the Balearic Islands (Spain). All subjects having started one of these 5 regimens including DTG with abacavir/lamivudine (ABC/3TC) or tenofovir fumarate/emtricitabine (TDF/FTC; regimens A and B, respectively), RAL with ABC/3TC (C) or TDF/FTC (D), or the co-formulation EVG/c/TDF/FTC (E) since July 2013 as their initial regimen or a switch with plasma HIV-1 RNA <50 c/mL were included. The incidence rate and 95% confidence interval [IR (95% CI)] of D/C due to toxicity is estimated as the ratio of the number of discontinuations by 100 patients/year of follow-up. Adjusted hazard ratios (aHR) and their 95% CI were obtained from multivariate Cox models, adjusted for gender, age, transmission group, origin, treatment-naïve and hepatitis B/C co-infection.

Out of 13066 patients on follow-up at July 2016, 2096 subjects were included (90% naives), receiving regimens A (n=859), B (n=108), C (n=208), D (n=280) and E (n=641). Of them, 430 (stopped prematurely their regimen, due to AEs in 74. The corresponding IR (95%CI) for DTG, RAL and EVG/c were 5.1 (3.6-7.0), 3.0 (1.8-4.5), and 2.8 (1.7-4.1), respectively. Among those receiving DTG, the IR with ABC/3TC or TDF/FTC were 4.9 (3.3-6.9) and 6.3 (2.0-12.9), respectively, with no significant differences between them. The aHR of D/C due to AEs with DTG vs. RAL was 1.1 (0.6-2.1), DTG vs. EVG/c 1.6 (0.8-2.9), and EVG/c vs. RAL 0.8 (0.4-1.6).

In subjects starting an initial therapy or a switch regimen with an undetectable plasma HIV-1-RNA, there are no significant differences in the D/C rates due to AEs among those receiving DTG, RAL or EVG/c. For those receiving DTG, there are no significant differences between those receiving ABC/3TC or TDF/FTC. Marginal structural models adjusted for baseline and time-varying confounding variables will be run in further analysis.

Session Number: 
Session Title: 
Presenting Author: 
Josep Llibre
Presenter Institution: 
Univ Hosp Germans Trias