Adverse Drug Reactions to Integrase Strand Transfer Inhibitors

Katherine J. Lepik; Benita Yip; Ana C. Ulloa; Lu Wang; Junine Toy; Linda Akagi; Viviane Dias Lima; Silvia Guillemi; Julio S.G. Montaner; Rolando Barrios


AIDS. 2018;32(7):903-912. 

In This Article


Study Cohort

There were 1522 new INSTI starts between 1 January 2012 and 31 December 2014. After excluding INSTIs initiated in clinical trials or outside British Columbia (n = 58), 1344 persons contributed 1464 unique person-INSTI exposures: 551 raltegravir, 394 elvitegravir-cobicistat, 519 dolutegravir. There were 110 persons who contributed data for two INSTIs and five persons for three INSTIs.

Baseline characteristics at INSTI start are summarized in Table 1. Overall, participants were predominantly male (79%) and ART-experienced (85%). There were more females and ART-naive persons in the elvitegravir-cobicistat group, more Hepatitis C coinfection in the raltegravir group, and more virologically suppressed ART switches in the dolutegravir group. Among ART-experienced persons, more than half had previously experienced an antiretroviral ADR, but few (<5%) had prior INSTI-related ADRs.

Within each INSTI category, baseline clinical and demographic characteristics were stable across time periods (data not shown). The proportion of persons with prior INSTI exposure increased according to the chronological introduction of each drug (Table 1). Few raltegravir-treated persons were INSTI-experienced, whereas nearly one-third of dolutegravir-treated persons had previously taken a different INSTI. Among the 1248 ART-experienced persons, 552 (44%) switched from protease inhibitor-based ART, and 362 (29%) from non-nucleoside reverse transcriptase inhibitor (NNRTI)-based ART (Supplemental Table S1, A total of 187 (15%) switched from a different INSTI, predominantly raltegravir changing to dolutegravir or elvitegravir-cobicistat. The most commonly stated reason for switching to the INSTI was an ADR to the previous regimen (approximately one-third of cases). ART simplification accounted for a high proportion of switches to dolutegravir and elvitegravir-cobicistat.

The majority (>98%) of persons received standard INSTI doses of raltegravir 400 mg twice daily, elvitegravir 150 mg daily, or dolutegravir 50 mg daily. Three raltegravir and 16 dolutegravir-treated persons received higher doses to overcome drug interactions or INSTI resistance (Supplemental Table S2,

At INSTI start, concurrent ART was most commonly two NtRTIs (76%, Table 1). Most (80%) 'other' ART combinations included a protease inhibitor (Supplemental Table S3, A higher proportion of raltegravir-treated persons received a background regimen other than two NtRTIs. Elvitegravir was available exclusively in combination with cobicistat, tenofovir disoproxil fumarate, and emtricitabine; therefore no elvitegravir-treated persons received an abacavir-lamivudine backbone. Other antiretroviral combinations with elvitegravir-cobicistat included addition of abacavir or a protease inhibitor (boosted by cobicistat in the combination tablet). During follow-up, the co-prescribed antiretrovirals were changed in 130 (8.9%) persons. The most common changes were switches between tenofovir disoproxil fumarate and abacavir (Supplemental Table S4, Tenofovir alafenamide was not available during the study period.

Follow-up and Censoring

Median follow-up for all INSTIs was 24 months (Table 2), with a cumulative follow-up of 2377 person-years INSTI exposure. In all groups, the majority of persons remained on the INSTI at the end of the 2-year follow-up, with retention highest for dolutegravir (81%) and lowest for raltegravir (58%). More raltegravir-treated persons discontinued treatment for non-ADR reasons (Table 2), with regimen simplification being the most common reason (Supplemental Table S5, There were also more deaths in the raltegravir group, but none were associated with INSTI ADRs.

Adverse Drug Reactions

Among 1464 person-INSTI exposures, there were 107 ADR reports implicating the INSTI. After excluding nine ADRs that did not result in INSTI discontinuation and one unclassifiable ADR, the remaining 97 ADR reports underwent blinded causality assessment. Eight reports were evaluated as unlikely related to the INSTI, leaving 89 study-defined ADR events. Reasons for assessment as 'unlikely' included the following: symptom onset preceded INSTI start, symptoms did not improve following INSTI discontinuation, symptoms did not recur with INSTI restart, and/or other, more plausible causes of the symptoms were identified.

Adverse drug reaction incidence was highest for elvitegravir-cobicistat-treated persons, with 38 of 394 (9.6%) having an ADR-related discontinuation compared to dolutegravir 27 of 519 (5.2%) or raltegravir 24 of 551 (4.4%) (Table 2). Unadjusted ADR rates (95% CI) were also higher for elvitegravir-cobicistat [5.94 (4.32, 8.16) per 100 person-years] than for raltegravir [2.91 (1.95, 4.35)] or dolutegravir [2.96 (2.03, 4.31)]. Most ADRs occurred in persons who received standard INSTI doses (88/89, 99%), and who remained on the same co-prescribed antiretrovirals throughout follow-up (87/89, 98%). For the two persons who changed co-prescribed antiretrovirals, the INSTI ADR was not temporally associated with the other antiretroviral alterations.

As displayed in Figure 1a, the higher incidence of elvitegravir-cobicistat ADRs was driven by a significantly higher proportion of persons with gastrointestinal ADRs (nausea, vomiting, diarrhea) 5.3% and general symptoms (fatigue, malaise) 1.3%. Central nervous system (CNS) symptoms such as sleep disturbances and headache were reported with all three INSTIs, and, although a higher proportion of dolutegravir-treated persons experienced a CNS ADR (3.5%, compared to elvitegravir-cobicistat 2.8% and raltegravir 1.6%), there was no statistically significant difference in CNS symptoms between INSTIs. Most CNS ADRs occurred in persons who received standard INSTI doses, with one case of sleep disturbance in a person treated with dolutegravir 50 mg twice daily. Details of all ADR symptoms are summarized in Supplemental Table S6-a (, and the 11/89 (12%) ADRs designated as 'severe' reactions by the reporter are summarized in Table S6-b.

Figure 1.

Adverse drug reactions (ADRs) to integrase strand transfer inhibitors (INSTI) by symptom class. ADRs leading to INSTI discontinuation (A) within 2 years and (B) more than 1 (up to 2 years) after INSTI start. *P<0.05 Pearson's chi-square or Fisher's exact test.

Symptoms associated with late ADR-related INSTI discontinuation (Figure 1b) were similar to the overall pattern, with significantly more gastrointestinal complaints in the elvitegravir-cobicistat group. Although time of symptom onset was not consistently documented, several cases of late ADRs did specify long-standing, suspected side effects prior to INSTI discontinuation.

The median time to ADR-related INSTI discontinuation was 3 to 4 months (Table 2). As depicted in the Kaplan–Meier plot (Figure 2) and Table 2, ADR-related discontinuations were reported throughout the 2-year follow-up period, but the majority (81%) occurred within the first year. Second-year INSTI discontinuations due to ADR were disproportionately higher in the elvitegravir-cobicistat group and accounted for (12/17, 70%) of late ADRs.

Figure 2.

Kaplan–Meier plot of cumulative probability of integrase strand transfer inhibitor (INSTI) discontinuation due to adverse drug reaction (ADR) during the first 2 years of treatment.

Confounder Model

In bivariate analyses, only INSTI drug and a prior history of antiretroviral ADR were significantly associated with INSTI ADR (Supplemental Table S7, In the adjusted Cox regression model (Table 3), the hazard ratio (95% CI) for elvitegravir-cobicistat versus dolutegravir ADR was 2.24 [(1.13, 4.44), P = 0.020]; however, there were no statistically significant differences in the ADR HR for either dolutegravir or elvitegravir-cobicistat versus raltegravir. A sensitivity analysis limited to period 3, when all three INSTIs were available, provided a similar result: adjusted ADR HR for elvitegravir-cobicistat versus dolutegravir 2.01 [(1.02, 3.98), P = 0.044].

Other Outcomes

Following ADR-related INSTI discontinuation, 22 of 89 (25%) of persons had documented ART interruption of several days to months, half of whom had virologic rebound above 1000 copies/ml. Whether directly switching to another regimen or re-starting ART after a period of interruption, 33 (37%) persons reverted to their previous ART regimen, 26 (29%) switched to a different INSTI, and 27 (30%) switched to a different non-INSTI regimen.

Selection of the subsequent ART regimen appeared to be influenced by the available antiretrovirals at the time of discontinuation. Only three of 20 (15%) of persons who stopped raltegravir or elvitegravir-cobicistat due to ADRs in periods 1 and 2 had a different INSTI in their next regimen, whereas 23 of 69 (33%) of persons with an ADR-related stop in period 3 switched to another INSTI, predominantly dolutegravir. Examination of the timing of ADR-related INSTI discontinuations did not reveal any clustering of either early or late ADRs around the market release of the newer drugs (Supplemental Figure S1, Most persons (70%) with a late INSTI ADR had started the INSTI in period 3, when all three INSTIs were available, and therefore had the opportunity to switch between INSTIs at any time.

There was no apparent relationship between the tolerability of one INSTI and subsequent tolerability of another INSTI. Among the 187 persons who had switched from a prior, tolerated, INSTI regimen to a new (study cohort) INSTI, 17 of 187 (9.1%) subsequently discontinued the new INSTI because of an ADR. The majority 15 of 17 (88%) of these ADRs occurred in persons who switched from raltegravir to either elvitegravir-cobicistat or dolutegravir for simplification. Conversely, among the persons with a study-defined INSTI ADR, 21 of 26 (78%) of those who switched to a different INSTI tolerated the new regimen (predominantly elvitegravir-cobicistat to dolutegravir switches). Two persons in the study cohort experienced more than one INSTI ADR: one developed a rash with both raltegravir and dolutegravir, and the other experienced elvitegravir-cobicistat related renal effects and dolutegravir CNS symptoms. Among the 89 ADR events, 10 persons (11%) eventually re-started the same INSTI due to patient or prescriber preference relative to other ART options.