Thank you for your interest in CCO content. As a guest, please complete the following information fields. These data help ensure our continued delivery of impactful education.
Become a member (or login)? Member benefits include accreditation certificates, downloadable slides, and decision support tools.
The first patient case is a 34-year-old black man who has sex with men (MSM). He was diagnosed with HIV 1 month ago and is treatment naive. He is a CrossFit trainer and nonsmoker who occasionally uses marijuana and alcohol, has a body mass index (BMI) of 22.5, and is ready to initiate ART. The patient has heard of and read articles about PWH experiencing weight gain with ART, and he asks to be started on a regimen “that won’t cause me to gain excessive weight.”
His baseline laboratory assessments are: CD4+ cell count, 523 cells/mm3; HIV-1 RNA, 187,000 copies/mL; negative for HLA-B*5701. His nonfasting glucose and creatinine clearance values were within normal range. He is immune to hepatitis B virus and is negative for hepatitis C antibody.
This very common scenario we encounter in clinical practice─patients newly diagnosed with HIV infection and no comorbidities─poses an important question: Should we take weight gain into account when deciding which regimen to recommend?
First, we will review the recommendations on first-line ART regimens from the DHHS and IAS-USA. Note that all of the regimens recommended by DHHS and IAS-USA include an INSTI, most commonly BIC or DTG.[1,2]
The 2-drug regimen DTG/lamivudine (3TC), which was approved by the FDA for first-line ART in April 2019, is now included as a regimen recommended by the DHHS for most PWH. However, all other recommended regimens include 2 different NRTIs, predominantly TAF or TDF. Thus, most patients initiating ART will be started on a regimen including an INSTI and/or TAF. This is important to bear in mind as we review data on weight gain associated with specific ARVs.
In 2019, Sax and colleagues published a pooled analysis in Clinical Infectious Diseases assessing weight gain in 5680 treatment-naive PWH drawn from 8 randomized phase III clinical trials between 2003 and 2015. The strengths of this analysis include a large study population that was quite diverse from clinical trials occurring over a long period of 12 years. In addition, each trial assessed weight up to 96 weeks following treatment initiation.
The pooled analysis reported that patients initiating ART gained a median of 2.0 kg by Week 96. The notable finding highlighted by these graphs is that the extent of weight gain differs by certain demographic characteristics. By Week 96, women were significantly more likely to gain weight than men; similarly, black PWH were significantly more likely to gain weight than nonblack PWH (both P < .05). When stratified by sex and race, the analysis showed that black women gained the greatest amount of weight (P < .05 vs nonblack women and P < .05 vs black men). By contrast, nonblack men and women gained the least amount of weight by Week 96.
Based on these data, we should expect that patients who start ART will gain weight. However, the likelihood and extent of the weight gain after starting ART is related as much to the patient’s characteristics as it is to the drug.
The pooled analysis further evaluated weight gain up to Week 96 by ARV class. Patients who started on INSTIs gained significantly more weight than those who started on a PI-based or NNRTI-based regimen (both P ≤ .05). By Week 96, the least squares mean weight change with INSTIs was approximately 3 kg compared with between 1 and 2 kg with a PI or NNRTI.
When weight gain was evaluated by INSTI (BIC, DTG, or cobicistat [COBI]-boosted elvitegravir [EVG]), the weight gained with BIC was similar to that with DTG. However, both of these INSTIs were associated with significantly higher weight gain vs EVG/COBI (both P ≤ .05).
These findings on weight gain differences among INSTIs have been consistent across studies. For example, my colleagues and I recently published a study in Clinical Infectious Diseases based on a cohort of ART-naive PWH at the Vanderbilt Comprehensive Care Clinic. In this cohort, we found less weight gain in those initiating EVG/COBI vs DTG. We also reported the least amount of weight gain in those initiating regimens based on EVG vs other INSTIs in an analysis of more than 22,000 treatment-naive participants in the NA-ACCORD study. However, it remains unclear why weight gain with EVG/COBI differs from other INSTIs and if the difference is driven by EVG or COBI.
This meta-analysis also demonstrates several interesting findings related to NRTI backbone. First, it is evident that weight gain was approximately 2 kg more in those starting TAF-based vs TDF-based regimens. Second, the weight gained with TAF, TDF, and abacavir was significantly higher vs zidovudine (all P ≤ .05).
These findings indicate that the NRTI backbone also plays a role in determining the extent of weight gain. As discussed earlier, most recommended regimens contain either BIC or DTG and TAF—meaning that PWH receiving recommended ART are thus on 2 ARVs associated with greater weight gain compared with other regimens.
The BMI of PWH has been progressively increasing over time. In this study presented at AIDS 2020, Silverberg and colleagues compared the BMI in PWH vs those without HIV in the Kaiser Permanente electronic medical records database (N = 138,222). The PWH had initiated treatment between 2006 and 2016. The uninfected controls were matched 1:10 by age, sex, race/ethnicity, clinic, and year. The investigators used linear mixed effect modeling to compare the changes in BMI over time in PWH vs uninfected controls.
Although both the PWH and uninfected controls experienced an increase in BMI, the rate of BMI increase in PWH was significantly greater at more than 3 times that of the uninfected controls (0.22 vs 0.06 kg/m2/year, respectively; P < .001). After 12 years, the average BMI was comparable between those with vs without HIV infection (28.4 vs 29.4 kg/m2, respectively). Indeed, the average BMI was in the higher end of the overweight category.
These data further support observations that the vast majority of modern PWH are no longer the malnourished, underweight patients who were starting ART in the 1990s and early 2000s. For example, a 2015 analysis drawing data from the Cross-Sectional Medical Monitoring Project and the National Health and Nutrition Examination Survey reported an obesity prevalence of 19% among men and 42% of women with HIV infection in medical care in the United States. This is concerning because additional weight gain increases this population’s risk for developing related cardiac and metabolic complications.
Continuing our discussion of the study by Silverberg and colleagues, PWH with a normal/underweight or overweight BMI at baseline gained weight at significantly higher rates than the uninfected controls (both P < .001). However, there was no significant difference in weight gain for PWH vs uninfected controls in the obese category (P = .09).
The take-home messages from this study are 1) that PWH have a high prevalence of BMI > 25 at baseline, and 2) PWH are at higher risk of gaining weight at a significantly greater rate than persons who are HIV negative.Regarding real-world patients, there is some amount of weight gain to be expected when starting ART, independent of the specific regimen. Previously, this was thought to be the “return to health” phenomenon. However, now that we are starting PWH on ART much earlier in their disease course, we are still witnessing weight gain after ART initiation, and the reasons for this weight gain remain to be determined.
Weight gain was also assessed in the multicenter, open-label phase III ADVANCE trial, which compared DTG plus FTC/TAF vs DTG plus FTC/TDF vs EFV/FTC/TDF in South Africa.[9-11] ADVANCE was one of the earliest randomized, controlled trials that compared both an INSTI vs an NNRTI and, at the same time, TDF vs TAF. This trial recruited ART-naive PWH who had not received ART in the past 6 months and who were neither coinfected with tuberculosis nor pregnant.
The primary endpoint was the proportion with HIV-1 RNA < 50 copies/mL at Week 48. The primary analysis found that DTG, whether given with FTC/TAF or FTC/TDF, was noninferior to EFV/FTC/TDF at Week 48, with virologic suppression rates of 84% vs 85% vs 79%, respectively. At Week 96, the rates of virologic suppression continued to be very similar at 79% vs 78% vs 74%, respectively, with no statistically significant differences found between the treatment arms.
This study also assessed weight gain as a secondary endpoint, and these results were very interesting and even a little surprising.
In both men and women, the weight gain with both DTG-based regimens was significantly higher than with EFV/FTC/TDF at Week 96 (all P < .05). For example, in men, those who were receiving DTG plus FTC/TAF had gained 5 kg, which was not significantly different from the 4 kg gained in those started on DTG plus FTC/TDF; however, these arms had both gained significantly more than the 1 kg observed in those started on EFV/FTC/TDF (both P < .01).
The more striking findings were among women. By Week 96, women started on DTG plus FTC/TAF had gained a mean of 10 kg, which was significantly higher than the 5 kg gained with DTG plus FTC/TDF and the 3 kg gained with EFV/FTC/TDF (both P < .001). Thus, among South African women, starting a DTG-based regimen was associated with significantly higher amounts of weight gain compared with starting an EFV-based regimen. Furthermore, starting a DTG regimen incorporating TAF was associated with significantly higher amount of weight gain than receiving DTG with TDF.
Of course, this might not be representative of all African women. The BMI for South African women tends to be higher than that of women elsewhere on the African continent. Nonetheless, the magnitude of this weight gain is remarkable.
At AIDS 2020, the ADVANCE investigators presented results on weight gain and metabolic syndrome through Week 144, although the data after Week 96 were incomplete. The findings were very similar to those from the earlier analyses. Women showed a trend of gaining more weight on FTC/TAF compared with those on FTC/TDF. Between Week 96 and Week 144, women receiving DTG plus FTC/TAF gained approximately 4 kg compared to approximately 3 kg with DTG plus FTC/TDF; those receiving EFV/FTC/TDF gained approximately 2 kg during this period. With men, the weight gain between Week 96 and Week 144 was not as pronounced as that in women.
Furthermore, most of this gained weight was fat mass rather than lean mass, and women gained significantly more fat mass than men (P < .001).
Particularly interesting was the investigators’ evaluation of cardiometabolic adverse outcomes. At Week 96, the rate of treatment-emergent metabolic syndrome was significantly higher for patients who were started on DTG plus FTC/TAF compared with those started on EFV/FTC/TDF (8.4% vs 3.9%, respectively; P = .03). It is striking that by Week 96, a notable proportion of patients who started on DTG plus FTC/TAF were developing metabolic symptoms.
The take-away messages from the ADVANCE analyses are 1) DTG-based regimens are associated with significantly greater weight gain than EFV-based regimens, 2) weight gain, particularly among women, is significantly higher with TAF and DTG than with either EFV-based ART or DTG with TDF, and 3) weight gain did not plateau but rather continued past 96 weeks. Finally, in addition to weight gain, there is also a significantly higher risk of developing metabolic syndrome for patients initiating DTG plus FTC/TAF.
An interesting post hoc analysis of the ADVANCE trial presented at CROI 2020 by Griesel and colleagues looked at the relationship between CYP2B6 genotype and weight gain with EFV. The investigators looked at 3 clinically relevant SNPs within CYP2B6 that affect EFV metabolism. Patients who were categorized as extensive metabolizers had lower blood levels of EFV compared with intermediate and slow metabolizers.
In this analysis, extensive metabolizers gained an amount of weight comparable to PWH who were receiving DTG/FTC/TDF (P = 1.00). Those who were slow metabolizers—and therefore had high blood levels of EFV—had significantly less weight gain compared with those receiving DTG/FTC/TDF (P < .001). Thus, the concentration of EFV in blood inversely correlates with the amount of weight gained.
Similar results had been presented earlier in PWH who switched from EFV-based to INSTI-based regimens, with the greatest weight gain noted in those who were slow metabolizers. When the PWH switched to an INSTI, they lost the weight-mitigating effects of the high concentration of EFV in their blood.
Returning to our discussion of key data from AIDS 2020, Chow and colleagues presented an analysis of ART-naive adults from the Decision Resources Group’s Real World Data repository who submitted a claim to start either a PI-based or INSTI-based regimen after July 17, 2018. They chose July 17, 2018, because that is the date that the FDA approved darunavir (DRV)/COBI/FTC/TAF. Among those who initiated PI-based regimens, almost two thirds were on the DRV-based single-tablet regimen, whereas approximately equal proportions of patients started either atazanavir-based or other DRV-based regimens. Among those who started an INSTI-based regimen, nearly 60% started a regimen based on BIC compared with 24% with DTG and 15% with EVG.
The investigators assessed the proportion of patients who had ≥ 5% or ≥ 10% weight gain, along with ≥ 5% or ≥ 10% BMI increase. Those receiving INSTI-based regimens trended toward a higher odds of being in all 4 of those categories. This is another real-world example demonstrating that PWH initiating INSTI-based ART are more likely to gain a significant amount of weight compared with those initiating PI-based ART.
Moving on to the OPERA study, this is a longitudinal prospective cohort analysis that used routine electronic health record data collected from approximately 8% of US PWH in care, which included more than 115,000 individuals drawn from 65 US cities and Puerto Rico. The investigators studied adults who were receiving TDF-containing 3-drug ART at baseline, who had ≥ 2 consecutive HIV-1 RNA assessments at < 200 copies/mL, and who switched from TDF to TAF. This is a common clinical scenario because we had many patients who were receiving TDF-based regimens until the FDA approved TAF-based regimens in 2015, after which many patients switched to TAF.
In this longitudinal cohort analysis, the investigators assessed weight changes predicted by linear mixed models adjusted for several clinical and demographic variables, including endocrine disorders and concomitant medications that could affect weight. The analysis included more than 3000 patients on INSTIs, most of whom were receiving EVG/COBI; approximately 1400 patients receiving NNRTIs, most on rilpivirine; and approximately 700 patients receiving boosted PIs, predominantly DRV.
The OPERA analysis found that, in those who maintained their other ARVs during the switch from TDF to TAF, there was a transient period of rapid weight gain between 0 and 9 months after the switch. During this time, PWH who switched from an INSTI and TDF to the same INSTI with TAF gained an estimated 2.64 kg/year; for those who maintained an NNRTI, they gained an estimated 2.25 kg/year; and for those who maintained a boosted PI, they gained an estimated 1.98 kg/year after the switch. The rate of weight gain plateaued across the different anchor ARV classes after 9 months.
Why is this important? Here, the analysis isolated the effect of switching the tenofovir formulation. The rapid weight gain in this short period of time and the consistency of the pattern across anchor ARV classes suggests that TAF has an independent effect on weight.
The OPERA investigators further examined weight change by individual INSTI when switching from TDF to TAF. Again, there was rapid weight gain between 0 and 9 months with EVG/COBI, DTG, and RAL. Those who were receiving RAL and TDF gained the least amount of weight when they switched to TAF while maintaining RAL. Nonetheless, the magnitude of the weight gain was comparable across the different INSTIs.
This model looked at weight changes in patients who switched from TDF to TAF while also switching to an INSTI—essentially 2 variables changing at the same time. During the period between 0 and 9 months, the extent of the weight gain differed somewhat. Indeed, those switching from TDF with no INSTI to TAF with BIC had an estimated weight gain of 4.47 kg/year between 0 and 9 months compared with 3.09 kg/year when switching to TAF and DTG, and finally 2.55 kg/year for TAF and EVG/COBI.
This large, retrospective cohort analysis indicates that in virologically suppressed patients switching from TDF to TAF, there appears to be an immediate increase in weight regardless of the concomitant ARVs being maintained. The weight is gained soon after the switch, within 0 and 9 months, and then plateaus across different regimens.
Moving on, van Wyk and colleagues evaluated metabolic changes in virologically suppressed adults switching from a 3-drug or 4-drug, TAF-based regimen to the 2-drug regimen DTG/3TC in TANGO, a randomized, open-label phase III noninferiority study. This was a very interesting analysis of the weight and metabolic effects when patients switch to a tenofovir-sparing regimen.
At Week 48, patients who switched to DTG/3TC gained 0.81 kg compared with 0.76 kg in those who maintained their TAF-based regimen. Therefore, there was no notable difference in weight gain after 48 weeks.
The analysis also evaluated whether this switch to a tenofovir-sparing regimen was associated with changes in the rates of insulin resistance, as measured by Homeostatic Model Assessment of Insulin Resistance, and metabolic syndrome. To summarize the data in the table, patients who switched to DTG/3TC vs those who continued TAF had similar median A1C values. However, the switch group had a lower rate of insulin resistance (65% vs 74%).
The investigators noted that PWH who switched from a boosted TAF-based ART regimen to DTG/3TC had a trend toward lower mean fasting glucose, a lower rate of insulin resistance, and a comparable rate of metabolic syndrome compared with those who maintained the boosted TAF-based ART. In the unboosted subgroups, I want to call attention to the difference in rates of metabolic syndrome between those who switched to DTG/3TC vs those who continued their TAF-based regimen. Among those who switched, 14% had metabolic syndrome at Week 48, which is an absolute 10% lower than with TAF-based ART. This difference was not statistically significant but was still notable.
Regarding the likelihood of developing insulin resistance, the adjusted odds ratio was significantly lower overall and in the boosted subgroup for those who switched to DTG/3TC vs continued their TAF-based regimen (P = .008 and P = .012, respectively). Although those who switched from an unboosted TAF-based regimen to DTC/3TC had a trend toward decreased odds of developing insulin resistance or metabolic syndrome, this decrease was not statistically significant.
Based on these results, it appears that PWH who switch off TAF will lower their likelihood of developing insulin resistance and metabolic syndrome. The odds of developing insulin resistance are significantly lower in those who switch to DTG/3TC, both overall and in the subset switching from boosted regimens.
An interesting finding from this analysis was the effects of the switch on lipids. Recall that TAF-based regimens have been previously associated with increased levels of low-density lipoprotein (LDL) and total cholesterol. In this analysis of TANGO, the investigators found that those patients who switched to a tenofovir-sparing regimen had significantly smaller changes from baseline to Week 48 in total cholesterol, LDL, triglycerides, and total cholesterol:high-density lipoprotein (HDL) ratio (all P < .05).
Looking at the boosted subgroup, switching from a boosted TAF-based regimen to a tenofovir-sparing regimen was associated with greater decreases in total cholesterol, LDL, triglycerides, and the total cholesterol:HDL ratio. This could be due to the effect of the boosting agent on these parameters and, because those patients were probably also receiving a PI, the association between PIs and an adverse lipid profile.
To summarize, this TANGO analysis indicates that among patients virologically suppressed on a TAF-based regimen, switching to DTG/3TC is accompanied by improvement in some metabolic parameters compared with continuing TAF-based ART. In the subset of those receiving a boosted TAF–based regimen, the lipid changes are even more favorable for those who switched to DTG/3TC.
How would I apply this in the clinic? I would consider switching to DTG/3TC for a patient who is virologically suppressed on a boosted TAF–containing regimen but who is prediabetic, at a high risk of developing diabetes, obese, and/or with hyperlipidemia. I would particularly consider switching for patients on a boosted regimen who have a preexisting cardiometabolic condition that puts them at higher risk for metabolic complications.
A post hoc analysis on weight and BMI changes was performed in DRIVE-SHIFT, an international, randomized, open-label phase III noninferiority trial.[23,24] This trial enrolled patients who were virologically suppressed on stable ART with normal kidney function and no evidence of prior virologic failure or resistance to study drugs. The participants were randomized to either switch immediately to doravirine (DOR)/3TC/TDF or to switch to DOR/3TC/TDF after 24 weeks.
As background, DOR is a next-generation NNRTI administered once daily that has several advantages compared with other NNRTIs: It has no major drug–drug interactions, tends to be well tolerated, and has fewer central nervous system adverse events compared with, for example, EFV. DOR can be taken without regard to food and can be taken with proton pump inhibitors.
In the DRIVE-SHIFT post hoc analysis, the investigators aimed to assess how weight changed when virologically suppressed patients switched to DOR/3TC/TDF. Of note, DOR is coformulated with TDF, and as discussed earlier, TDF is associated with less weight gain than TAF. Furthermore, the weight gain associated with NNRTIs tends to be less than with PIs or INSTIs.
This graph illustrates that the weight gained after switching to a DOR-based regimen was modest, ranging from a mean of 0.7 kg to 1.4 kg at 24 and 144 weeks past the switch, respectively. This change is quite minimal compared with that discussed earlier when switching from a TDF-based to a TAF-based regimen.
Shown here are data on excessive weight gain and changes in BMI categories after switching to the DOR-based regimen. Of note, a ≥ 10% weight gain occurred in 2.5% to 2.6% of patients at 24 weeks after switching and in 7.9% to 8.4% at 144 weeks after switching. The mean weight change was less in those who switched from an INSTI vs a PI and NNRTI.
The majority of the patients with a normal, overweight, or obese BMI—between 80.2% and 95.5%—maintained their BMI category at Week 144 after switching. Of interest, 50% of those who were underweight in the immediate switch arm shifted to a normal BMI category at Week 144, whereas all underweight patients in the delayed switch arm remained underweight by BMI at Week 144.
These results indicate that in the large majority of PWH who have a normal or high BMI, switching to DOR does not increase their risk of experiencing obesity or being overweight. On the other hand, in PWH who are underweight, switching to DOR increases their likelihood of having a normal BMI after 144 weeks. I would emphasize that the weight changes were modest in these virologically suppressed PWH switching to a DOR-based regimen.
How should we translate this to the clinic? Let’s take as an example an ART-naive patient who started BIC/FTC/TAF and gained a notable amount of weight. Should we now switch that patient to DOR/3TC/TDF to try and reverse that weight gain? Based on this study, I would say no because there is still some weight gain after switching to DOR.
Several other interesting studies on weight gain reported at AIDS 2020 included the AFRICOS study, which assessed adult PWH who were starting or switching ART at 12 different PEPFAR sites in Uganda, Kenya, Tanzania, and Nigeria. In AFRICOS, the likelihood of developing a high BMI with DTG/3TC/TDF was 1.85 times higher compared with other regimens. There were no notable differences in hyperglycemia incidence. This is quite relevant to low- and middle-income countries such as Kenya, where PWH are now being started on DTG/3TC/TDF rather than older regimens such as EFV/3TC/TDF.
Another interesting study from AIDS 2020 was a pooled analysis of virologically suppressed adults aged 65 years or older who switched to BIC/FTC/TAF through clinical trials. The median weight gain was 1.0 kg at Week 48, indicating that even among older PWH, switching to an INSTI-based regimen with TAF is associated with some degree of weight gain.
Having reviewed the data suggesting that PWH receiving TAF are at higher risk of experiencing weight gain compared with those receiving TDF, we can turn to this retrospective cohort study investigating whether PWH receiving TAF are also at higher risk of developing cardiac events. As shown in the left graph depicting an unadjusted Kaplan-Meier model of cardiac event-free survival, PWH receiving TAF had the shortest duration of cardiac event-free survival compared with those receiving TDF or tenofovir-sparing regimens.
After adjusting for other clinical and demographic variables, the investigators reported a significantly higher risk for cardiac events among those receiving TAF compared with those receiving tenofovir-sparing ART (adjusted HR: 3.86; 95% CI: 1.51-9.82; P = .005). Similarly, the risk of cardiac events was significantly higher for those receiving TAF compared with those receiving TDF (HR: 1.94; 95% CI: 1.05-3.57; P = .034).
This is one of the first studies to assess whether the greater weight gain associated with TAF is also associated with clinical complications, namely, the risk of cardiac events.
The DHHS guidelines recognize weight gain as a common adverse event associated with ART initiation and subsequent virologic suppression. The increase is greater with INSTIs vs other ARV classes, with TAF vs TDF, and with DOR vs EFV. However, there are no recommendations at this time to avoid INSTIs or TAF out of concern for potential weight gain.
I always tell my patients that we still do not know why greater weight gain is noted with INSTIs compared with, for example, NNRTIs. Is it an off-target effect of the drug itself, or is it because these INSTIs have greater efficacy, are better tolerated, and are decreasing the energy expenditure and level of inflammation within adipose tissue and thereby further decreasing the catabolic state? It remains unclear. Although this weight gain might be associated with a higher risk of metabolic complications, the mechanism has not been established yet.
Returning to the case presented at the beginning of our discussion on ART and weight gain, recall that the patient was a 34-year-old black MSM and a CrossFit trainer newly diagnosed with HIV who asked to be started on a regimen “that won’t cause me to gain excessive weight.”
For my treatment-naive patients ready to start ART, I would not change my choice of treatment to avoid TAF or BIC. If I think that, for example, BIC/FTC/TAF is the most appropriate regimen to use for that patient, I would use it. I would caution that there is some associated risk of weight gain, but we do not know why—whether the weight gain is because the drugs are more effective or some other mechanism.
I also note that weight gain is likely to happen with all ART regimens but to the least extent with EFV. EFV has gastrointestinal and central nervous system adverse events, and as we discussed earlier, weight gain is less in patients who are slow metabolizers and have high EFV levels in their blood. Taken together, this may mean that even a regimen associated with less weight gain is not necessarily the better treatment choice.
Until we have a better understanding of both the cause of and the long-term complications from this weight gain, I will continue to adhere to the DHHS guidelines for treatment-naive PWH.
There were multiple take-home points from recent data on ART and weight gain.
We need more studies to understand this phenomenon of higher weight gain with TAF vs TDF, and to elucidate why we see different effects on the lipid profile with these 2 tenofovir formulations.
You are accessing CCO's educational content today as a Guest user.
If you would like to continue with free, full access to the CCO Web sites, including free CME/CE credits, please click the button below.