Do Integrase Inhibitors Cause Weight Gain?

(See the Major Article by Bourgi et al on pages 1267–74.) A majority of initial and salvage antiretroviral (ARV) regimens for human immunodeficiency virus (HIV) treatment now contain an integrase strand transfer inhibitor (INSTI). The reason: relatively high tolerability and greater likelihood of sustained treatment success compared to other classes [1]. The INSTI dolutegravir (DTG), in particular, revolutionized antiretroviral therapy (ART). Due to a remarkably high barrier to resistance and other advantages, it quickly became the go-to ARV anchor drug in the United States following approval in 2013. However, practitioners must now incorporate several postmarketing safety signals into clinical decision making. Reports of central nervous system (CNS) side effects with DTG were not shocking because we had all seen individuals quit the medicine due to headache or insomnia. A warning about neural tube defects was startling but has not been confirmed. Now, another surprising postapproval observation has evolved into an overt clinical conundrum: do INSTIs, especially DTG, lead to excessive weight gain? If so, does this signify toxicity or faster return to health? A fascinating story has built to this predicament. In 2017, a short research letter described weight gain as an “unexpected bothering side effect” of DTG [2]. Authors observed that 55 of 517 (11%) individuals prescribed the drug at a single center (most of whom had switched from an alternate agent) stopped it due to intolerability; 4 of those 55 (7%) reported weight gain as the intolerable side effect. Weight increases while on DTG ranged from 4 to 12 kg. This report raised eyebrows but did not cause much of a stir. Shortly thereafter, 2 larger retrospective studies corroborated the association between INSTIs and weight gain. In one, a switch from efavirenz (EFV) to an INSTI led to more weight gain than continuation with EFV; weight change was most pronounced with DTG [3]. In the other study, initiation of an INSTI led to a greater body mass index (BMI) gain per year, higher obesity incidence, and reduced time to obesity compared to initiation of a boosted protease inhibitor (bPI) or nonnucleoside reverse transcriptase inhibitor (NNRTI) [4]. Although the number of patients who took an INSTI in this report was small (and most were taking raltegravir [RAL]), INSTI use was the strongest predictor of clinical obesity (increasing the risk 7-fold). These reports turned more heads, but still many wondered, is this real? Now, a barrage of cohort studies have substantiated that the findings are real and should not be ignored. This issue of Clinical Infectious Diseases highlights one such study. Bourgi et al performed a retrospective analysis of treatment-naive persons with HIV at a single center who started ART between 2007 and 2016 [5]. Controlling for sex, baseline BMI, CD4 count, and HIV RNA, adjusted average weight gain was greatest with DTG at 6 and 18 months (significantly greater than with NNRTIs and elvitegravir [ELV] and higher than with RAL and bPIs, though that difference did not reach statistical significance). There was a trend toward higher obesity incidence with DTG. While limitations abound (single center, somewhat small numbers taking each INSTI, mostly male cohort, relatively high prevalence of obesity at baseline), the study adds to what seems to have swelled overnight into an indisputable collection of evidence that shows that INSTIs, especially DTG, are associated with more weight gain than other ARVs. At the Conference on Retroviruses and Opportunistic Infections (CROI) 2019, at least 6 observational studies examined weight changes with INSTIs. In the largest of these (another retrospective analysis led by Dr. Bourgi), investigators examined weight change in more than 24 000 treatment-naive persons in North American AIDS Cohort Collaboration on Research and Design cohorts and found that at 5 years after ART initiation, INSTIs led to greater change in predicted weight compared to NNRTIs and bPIs (though the bPI difference was not statistically significant) [6]. At 2 years, DTG led to nonsignificantly more weight gain than RAL, and both led to significantly more weight gain than ELV. The 5 additional studies (1 prospective and 4 retrospective, including 1 in women only) examined weight changes in virologically suppressed, treatment-experienced persons who switched to or added an INSTI; 4 of 5 found significantly greater weight gain with INSTIs compared to non-INSTIs [7–11]. Data from randomized clinical trials (RCTs) have been summarized recently [12]. Though not designed to compare weight changes, RCTs that reported comparisons of weight substantiate the cohort study findings. Turning to critical questions: Do INSTIs lead to weight gain directly or indirectly? Does the exaggerated weight gain with INSTIs increase the risk of metabolic disturbances or cardiovascular events? Does it magnify the likelihood of neurocognitive impairment (since obesity and metabolic syndrome are risk factors)? In other words, is the weight gain a bothersome side effect or something more sinister? Answers remain unclear. Some studies identified specific risk factors for weight gain with INSTIs, including older age, female sex, and black or Hispanic ethnicity [8, 10]. The current Bourgi et al analysis found no difference between sex or race, but numbers of women were small [5]. Pharmacogenetics may predict weight changes with specific ARVs but are not fully understood. Per 1 retrospective investigation, CYPB26 EFV slow-metabolizers gained more weight than others after they switched to an INSTI; however, this finding varied based on racial characteristics and specific INSTI [13]. Why would EFV slow-metabolizers gain more weight after changing to an INSTI? It was suggested that slower EFV metabolism before switching correlates with feeling better after switching and thus consumption of more calories; however, no study to date has controlled for activity level or caloric intake. In addition, generalized weight gain seems to occur after the switch to an INSTI from a variety of preswitch agents, so this explanation seems far from complete [7, 10]. It has also been proposed that perhaps INSTIs lead to exaggerated weight gain due to faster viral load suppression. Perhaps this induces more rapid reduction of HIV-related inflammation and basal energy expenditure [4]? A compelling part of Bourgi et al’s data addresses this question [5]. Viral load reductions between specific INSTIs were not significantly different over the first 12 months, yet weight change differed. At 18 months, viral loads were the same across ART classes, but weight change varied. Thus, viral kinetics do not explain the weight findings. What other mechanisms could be contributing to this? Could direct off-target effects of INSTIs be to blame? Bourgi et al discuss data that indicate that DTG may increase appetite through stimulation of melanocyte-stimulating hormone [5]. This is captivating but not confirmatory. Some analyses have examined ARV effects on adipocyte differentiation and function (adipose tissue is not passive but part of a complex signaling network). These studies found that ELV affects adipocytes less than EFV and that RAL causes neutral effects on adipocytes, but this would not explain weight gain after switching EFV to an INSTI [14, 15]. To complicate matters further, a recent investigation observed improvements in insulin sensitivity after a switch from a bPI to DTG or RAL, despite decreases in serum leptin levels and a trend toward increased waist circumference with both INSTIs [16]. This is difficult to juxtapose with other INSTI weight change findings since actions of leptin (a hormone secreted by adipose cells) and mechanisms of leptin sensitivity are complex. Our understanding of the effects of ARVs on regulation of energy expenditure and storage remains limited. Do side effects like CNS intolerability or weight gain differ with newer INSTIs such as bictegravir (BIC) or investigational cabotegravir (CAB)? Anecdotally, I have not seen nearly the rates of headache or insomnia with BIC as found with DTG, suggesting some difference in CNS properties, though effects of BIC on weight have not been examined thoroughly. One RCT reported weight changes with BIC to be similar to those with DTG, but this must be confirmed [17]. Curiously, a study of intramuscular CAB for persons without HIV did not detect significant weight gain over time; however, whether this is due to differences in the drug, drug delivery, or host must be investigated [18]. Additional outstanding questions beg study. Does the NRTI backbone play a role? A recent single-center study suggested that tenofovir alafenamide (TAF) leads to more weight gain than tenofovir disoproxil fumarate (TDF) in the first year after initiation and that switching TDF to TAF is associated with weight gain, so this question should be addressed [19]. Bourgi et al lacked an adequate number of patients taking DTG with TDF or TAF to make an adequate comparison to abacavir [5]. Few of the cohort studies have tackled the NRTI question. Additionally, most (not all) studies found less weight gain with ELV compared to RAL or DTG, but is ELV unique or does the requisite cobicistat booster make the difference? Returning to the most crucial point of uncertainty: is weight gain with INSTIs altering clinical outcomes? Weight gain affects self-perceptions and may limit a person’s willingness to continue ART, but weight changes, especially increases in central adiposity, also raise risks of numerous comorbidities. However, data that compare weight-associated clinical outcomes are starkly limited. In the retrospective analysis of weight and body measurements in women with HIV, hypertension was more frequent after a switch to or addition of an INSTI compared to continuation on a non-INSTI, but clinical endpoints are overall lacking in the cohort studies [7]. The analysis by Bourgi et al adds pieces to the puzzle, but huge gaps remain. For now, clinicians must weigh the pros and cons of various INSTIs; however, I do not think practice or guidelines should change. Practitioners should incorporate counseling and shared decision making about potential weight change into ARV initiation and switch visits and monitor for metabolic complications, but INSTIs should remain as first-line treatment. Notably, in the Bourgi et al study, no person taking DTG, RAL, or ELV changed ART over 18 months, and HIV RNA levels were more likely to remain suppressed while on these compared to non-INSTI options, underscoring the enormous treatment advantage of INSTIs [5]. It is essential that future studies seek a better understanding of the mechanisms and long-term clinical implications of INSTI-associated weight gain and include comparisons of agents that now predominate HIV clinical care, such as BIC and TAF. Potential conflicts of interest. The author: No reported conflicts. The author has submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.