Treatment Outcomes of Integrase Inhibitors, Boosted Protease Inhibitors and Nonnucleoside Reverse Transcriptase Inhibitors in Antiretroviral-Naïve Persons Starting Treatment

A Mocroft; B Neesgard; R Zangerle; A Rieger; A Castagna; V Spagnuolo; A Antinori; FC Lampe; M Youle; JJ Vehreschild; C Mussini; V Borghi; J Begovac; C Duvivier; HF Gunthard; A Rauch; J Tiraboschi; N Chkhartishvili; N Bolokadze; F Wit; JC Wasmuth; S De Wit; C Necsoi; C Pradier; V Svedhem; C Stephan; K Petoumenos; H Garges; F Rogatto; L Peters; L Ryom

Disclosures

HIV Medicine. 2020;21(9):599-606. 

In This Article

Results

Of 5198 eligible ART-naïve persons in RESPOND, 2358 (45.4%) started INSTIs (1342 dolutegravir, 429 raltegravir, 580 elvitegravir and seven bictegravir), 1349 (26.0%) PI/bs (976 darunavir and 373 atazanavir) and 1491 (28.7%) NNRTIs (823 efavirenz and 668 rilpivirine). The majority were male (n = 4248; 81.7%), of white ethnicity (n = 3617; 69.6%), and men who have sex with men (n = 2908; 55.9%). The most commonly used nucleoside backbones were tenofovir disoproxil fumarate/emtricitabine (n = 3728; 71.7%) and abacavir/lamivudine (n = 925; 17.8%). Those starting PI/bs were more likely to be female, to have a higher VL and to have a lower nadir CD4 count. Those starting INSTIs had started ART more recently.

Overall, 4700 persons (90.4%) had 12 months of follow-up; 2762 [58.8%; 95% confidence interval (CI) 56.9–60.6%] achieved cTO success (Table 1). The proportion with cTO success was highest for INSTIs and NNRTIs and lowest for PI/bs (61.7%, 63.3% and 49.5%, respectively). After adjustment, those on PI/bs had lower odds of cTO success [adjusted odds ratio (aOR) 0.74; 95% CI 0.64–0.87] with no significant differences comparing INSTIs and NNRTIs. There was no evidence that the differences in cTO success between INSTIs, PI/bs and NNRTIs differed according to age group, CD4 count or VL at baseline, or according to the presence or absence of severe immunosuppression (all P-interactions > 0.1). A much higher proportion achieved virological success in the on-treatment analysis (96.8%; 95% CI 96.2–97.4%; Table 1 and Figure 1). After adjustment, those on PI/bs had lower odds of cTO success (aOR 0.44; 95% CI 0.22–0.89), with nonsignificant differences comparing INSTIs and NNRTIs. There was no evidence that the differences in on-treatment virological success between INSTIs, PI/bs and NNRTIs differed according to age group, CD4 count or VL at ART initiation, or according to the presence or absence of severe immunosuppression (all P-interactions > 0.1).

Figure 1.

Outcomes after initiation of antiretroviral therapy (ART) in ART-naïve persons. *cTO, immunologic and on-treatment outcomes using logistic regression and adjusted odds ratio; clinical outcomes using Poisson regression and adjusted incidence rate ratios. Models adjusted for cohort, CD4, age, viral load, AIDS, HBV, HCV, year of starting ART, gender, race, HIV risk, time since HIV diagnosis, baseline date and nucleoside backbone (all at starting ART).

In total, 3979 (76.5%) persons had 12 months of follow-up and a CD4 count measured at 12 months, and 905 (22.7%; 95% CI 21.4–24.0%) achieved immunological success (Table 1). The proportion with immunological success was highest for those starting INSTIs (Table 1). After adjustment, particularly for age and nadir CD4 count, those on NNRTIs had lower odds of immunological response (aOR 0.67; 95% CI 0.52–0.85), with no statistically significant differences comparing INSTIs and PI/bs (Figure 1). There was no evidence that the differences in immunological response between INSTIs, PI/bs and NNRTIs varied according to age group, CD4 count or VL at ART initiation, or presence or absence of severe immunosuppression (all P for interactions > 0.1).

A total of 258 persons had a new AIDS diagnosis or died > 14 days after ART initiation during 15 466 person-years of follow-up (PYFU) (incidence rate 16.7/1000 PYFU; 95% CI 14.6–18.7). Among those with an event, the median time to event was 4 months [interquartile range (IQR) 1–15 months]. The most common event was death (n = 63; 24.4%), followed by pulmonary tuberculosis (n = 28; 10.9%) and non-Hodgkin's lymphoma (n = 26; 10.1%). There were no differences in the proportion with VL > 200 copies/mL at event (132/258; 51.2%) across ART classes (P = 0.028). Those starting INSTIs had lower CD4 counts at events (median 139 cells/μL) compared to those starting NNRTIs (median 221 cells/μL) and PI/bs (median 212 cells/μL; P = 0.013). The crude incidence was highest for those starting INSTIs and lowest for those starting NNRTIs (Table 1). After adjustment, particularly for age and nadir CD4 count, there were no significant differences in the incidence of new AIDS diagnoses or death between ART classes (Figure 1). Importantly, this finding was consistent across age groups, CD4 counts and VLs at baseline, and between those with and without severe immune suppression (all P for interactions > 0.1).

Results were consistent for cTO and virological success using a lower limit of detection of 50 copies/mL, when immunological success was defined as a CD4 count increase to > 500 cells/μL (baseline CD4 count < 400 cells/μL) or a 25% increase in CD4 count (baseline CD4 count > 400 cells/μL), for cTO, virological and immunological success at 6 months after starting ART, and for clinical progression to new AIDS diagnosis/death censoring at first change to regimen started (data not shown).

Comments

3090D553-9492-4563-8681-AD288FA52ACE

processing....