ART, weight changes and obesity: results from CASCADE cohort

16 September 2024. Related: Side effects, Lipodystrophy and metabolic complications.

Simon Collins, HIV i-Base

A large retrospective study of weight changes associated with different HIV drugs in the international CASCADE cohort of people in very early infection, has been published in Lancet HIV together with an editorial on this complication of modern HIV therapy. [1, 2]

The study is important for looking at the impact of ART that has been initiated soon after HIV infection and in people who were using treatment for the first time. Looking into this time period minimises the impact of residual confounding from both potential metabolic complications related to chronic HIV and the potential impact of previous ART. The results contribute important data to one of the least explained issues experienced by many people living with HIV.

Enrolment criteria for this analysis included having started ART within a year of documented seroconversion, a record of pre-ART weight and BMI, and at least one set of post-ART blood samples before either first treatment change or the end of May 2023.

The study modelled the risk of average weight changes of >5% and >10% after 6 months and 3 years on ART by baseline BMI and ART class and adjusted for all key demographic and baseline factors including sex, gender, likely route of infection, weight, BMI category, calendar time and drug class (INSTI, PI/b and NNRTI).

The analysis includes data from 5698 participants in the CASCADE study who started ART between 2007 and 2022. CASCADE is a cohort of adults (16 and older) who were diagnosed during acute, primary or early infection at participating clinics in France, Greece, the Netherlands, Spain, Sweden, Canada and the UK. Approximately 10,000 potential participants in the cohort were not included because they either started ART more than a year after seroconversion or due to missing HIV or baseline data.

Median age was 33 (IQR: 26 to 43), time from diagnosis to ART was 4 months (IQR: 1.9 to 6.3), with baseline CD4 of 459 cells/mm³ (IQR: 328 to 620) and viral load of 4.9 copies/mL (IQR: 4.3 to 5.5).

It is important that the study was designed to capture data in transgender people. Two participants were transgender women who for this analysis were included with the MSM group. Overall, 90.3% were assigned male at birth, 9.1% female at birth and 0.6% had missing data. Approximately 80% of participants were gay or bisexual men and 75% were of European or North American origin, used as a marker for ethnicity.

ART was INSTI-based for 49%, boosted PI-based for 32% and NNRTI-based for 19% with choice significantly linked to baseline characteristics, including year of diagnosis, route of transmission, viral load etc, reflecting guidelines and clinical practice. For example, dolutegravir (42%) was the most widely prescribed INSTI, followed by elvitegravir (26%), bictegravir (20%) and raltegravir (11%). Tenofovir alafenamide (TAF) was prescribed for approximately 17% of participants, nearly always with an INSTI (89%) – mainly bictegravir or elvitegravir.

Results

Significant differences in weight were reported by ART, BMI, sex and ethnicity and the study results were detailed and complicated.

They included significant differences in weight gain by drug class and by individual drugs within class. INSTI-based ART showed the greatest average increase, following by PI/b and then NNRTI, with these differences seen in all BMI categories (p=0.011).

There were also significant differences by baseline BMI with weight gain generally inversely related to baseline BMI. Those with lowest BMI at baseline had the greatest weight gain and those with highest baseline BMI either gained less weight or lost weight.

For example, among people with normal BMI at baseline (18.5-24.9), after three years, weight increased by >10% in 30% of people using INSTI-, 25% using PI/b- and 20% of those using NNRTI-based ART. (See Table 1).

Within the same BMI band, average (IQR) weight gains after three years on INSTI-based ART were 5.63 kg (4.92 to 6.35 kg) in women, 5.76 kg (5.06 to 6.46) in people from sub-Saharan African, and 3.82 kg (3.50 to 4.13) in gay men. Increases were higher in women compared to men irrespective of drug class (p<0.0003).

In every BMI category, increases in people using INSTI-based ART were highest in people from sub-Saharan African compared to European gay men.

Examples of combinations with the highest and lowest estimated weight changes are shown in Table 2. These were calculated for European gay men age 30 to 39, who were the largest demographic group, based on average height and baseline CD4 and viral load.

Table 1: Percentage of participants (BMI 18.5 to 24.9) with >5% and >10% weight increase after 3 years ART*

Note: Figures rounded to nearest percent.

Table 2: Highest and lowest kg weight increases (95%CI) by baseline BMI*

*Estimated for European gay men age 30 to 39, with average height and baseline CD4 and viral load.

These results are important for showing independent associations between weight gain and ART in a large cohort, predominantly of gay European men, but that also support greater increases in women and in people from South Africa reported in the randomised NAMSAL study. However, weight gains in CASCADE were significantly lower than those reported in NAMSAL and also in the larger ADVANCE study.

Table 3: Average weight gain reported in the NAMSAL study

The CASCADE results are important for supporting a direct effect of INSTIs and tenofovir alafenamide on bodyweight gain, rather than this being a return to health effect, because participants were still in very early infection.

The researchers also concluded that weight management needs to be actively included in the routine management of ART that includes these drugs, because of the impact of future health, including cardiovascular disease.

The accompanying editorial by Jennifer Manne-Goehlera and Mark J Siedner discusses the difficulty of separating the impact of treatment in an observational study from potential confounding factors, and in distinguishing excessive weight gain from a return to health effect.

comment

These results contribute a new and unique dataset that minimises the impact of long-term HIV infection including advanced HIV. This potentially makes it easier to report the impact of different ART.

Many people still report that significant unexplained weight gain after starting ART as a highly distressing outcome and yet, more than five years after the NAMSAL and ADVANCE studies, guidelines are still not able to recommend effective management and how this can be reversed.

In addition to significantly reduced quality of life, the additional increased health concerns include a higher risk for diabetes and cardiovascular disease.

Weight changes, especially increases in weight, have been reported since 1997 when the first triple-combination therapies, then referred to as HAART, included lipodystrophy as one of the diverse metabolic side effects. The mechanism of weight and fat changes then was never explained and some researchers argued that weight increases broadly just reflected the background patterns of obesity in the general non-HIV population.

Data supporting the role of specific ARVs are challenged by the lack of weight loss after switching to other drugs.

One hypothesis not discussed in either the main paper or the editorial comment is whether the assault from seroconversion itself has the potential to cause metabolic changes that might only be unmasked much later in the course of infection, perhaps triggered by ART or individual drugs.

The multifactorial nature of excessive weight gain also suggests that multiple interventions should be included in trying to reduce weight, for example that studies switching away from TAF, perhaps from INSTI to doravirine, should also include appropriate diet and exercise and other interventions. [3]

The expected benefits from individualised diet and exercise programmes are often limited by short-term follow-up and difficulty in sustaining behavioural changes.

The potential benefits from using GLP-1 agonists such as semaglutide have been reported in people living with HIV in one small US study. An important caution is that these drugs reduce weight systemically and have the potential to worsen existing HIV-related lipoatrophy.

The ongoing SWIFT study using semaglutide is still enroling at sites in Dublin and Liverpool. [4]

References

1. Pantazis N et al for the CASCADE Collaboration. Changes in bodyweight after initiating antiretroviral therapy close to HIV-1 seroconversion: an international cohort collaboration. Lancet HIV. 23 August 2024.
   https://www.thelancet.com/journals/lanhiv/article/PIIS2352-3018(24)00183-8/fulltext
2. Manne-Goehler J and Sieder MJ. Untangling the causal ties between antiretrovirals and obesity. Lancet HIV, 23 August 2024.
   https://www.thelancet.com/journals/lanhiv/article/PIIS2352-3018(24)00214-5/fulltext
3. Mounzer K et al. Weight change after starting doravirine among ART-experienced individuals in the US. CROI 2024, Denver. Poster abstract 807.
   https://www.croiconference.org/abstract/weight-change-after-starting-doravirine-among-art-experienced-individuals-in-the-us/
4. ClinicalTrials.gov. Semaglutide’s efficacy in achieving weight loss for those with HIV (SWIFT).
   https://clinicaltrials.gov/study/NCT04174755