| |
No overall impact on rate of weight gain with integrase inhibitor-containing regimens in antiretroviral-naïve adults
|
| |
| |
Download the PDF here
Our real-world cohort was on average gaining weight both pre- and post-ART in our unadjusted analysis. However, we did not observe a difference in the rate of weight or BMI gain after starting an ART regimen containing RAL or DTG compared with a PI- or NNRTI-based regimen. Our study is novel in that we incorporated weights before baseline to assess pre-ART trends. An increased rate of weight gain was observed in those starting treatment with lower CD4, higher VL and older age, probably associated with a return to health. Both findings have been observed in other HIV cohorts [6].
Weight change after ART initiation is probably a complex, multifactorial process. These data suggest that RAL- and DTG-containing regimens in treatment-naïve individuals are not associated with an acceleration in weight gain compared with other ART regimens, beyond that expected following viral suppression, within the limitations specified. Research incorporating assessment of general weight gain factors such as diet and exercise would be beneficial, mitigating the under-reporting and recording bias with diet recall, although this is often logistically challenging [18]. Comorbidities also need to be considered; depression was associated with increased rate of weight gain, a condition with high prevalence [19] where both the disease and the treatment can affect weight [20].
Caution should be shown in attributing weight gain as a causal effect of INSTI use in the absence of adequate follow-up and monitoring. More data are needed to evaluate individuals switching away from INSTIs because of weight gain. This should include capturing obesogenic factors; whether the weight gain continues, plateaus or reverses; weight distribution, and whether any clinical or metabolic sequelae manifest.
Although ART-naïve individuals may experience gains in absolute weight after starting ART, the cause is likely to be multifactorial. One hypothesis to explain the variability of effect could be that INSTIs sensitize PLWH to gaining weight, the melanocortin-4 receptor pathway being one proposed mechanism [9]. However, absolute gains are probably influenced by lifestyle [10], cultural attitudes to weight [11], the individual's perception of their weight [12], and host genetics/microbiome [13]. This may explain why weight gain is not a universal phenomenon for all individuals on ART and for all regimens [14]. Improved tolerability of modern regimens could also be a factor and is more easily identified when full adverse event data from studies are reported [7]. For individuals who undergo extremely rapid weight gain with ART, host/genetic factors, yet to be elucidated, could amplify or synergize such an effect.
Individuals with a more recent year of ART initiation were heavier at initiation. This probably reflects the fact that PLWH are now diagnosed and started on treatment earlier, with fewer patients presenting with advanced disease and the associated weight loss. However, it may also reflect the increasingly obesogenic environment that equally affects the general population; 64% of adults in England are overweight or obese (36% and 28%, respectively) [15], with a similar prevalence observed in comparable countries [16]. An increased rate of weight gain was observed in those starting treatment with lower CD4, higher VL and older age, probably associated with a return to health. Both findings have been observed in other HIV cohorts [6].
---------------------------
No overall impact on rate of weight gain with integrase inhibitor-containing regimens in antiretroviral-naïve adults
11 October 2021
James E. Burns1,2 | Oliver Stirrup1 | Laura Waters2 | David Dunn1,3 | Richard Gilson1,2 | Sarah L. Pett1,2,3
Abstract
Objectives
Integrase strand transfer inhibitors (INSTIs) are commonplace in modern antiretroviral therapy (ART). Increased weight gain with their use is increasingly scrutinized. We evaluated weight changes in treatment-naïve adults with HIV-1 attending a UK centre who started regimens including raltegravir or dolutegravir.
Methods
A retrospective cohort study of adults prescribed an INSTI between January 2015 and March 2020 were categorized as having started an ART regimen containing raltegravir, dolutegravir, a protease inhibitor or a nonnucleoside reverse transcriptase inhibitor. Individuals with one or more weight measurement ≤ 5 years both pre- and post-ART initiation, who started a three-drug regimen with ≥ 6 months duration and achieved virological suppression (< 50 copies/mL) within 6 months were included. A random effects model with linear slope pre- and post-ART was used, adjusting for age, gender, ethnicity, ART regimen, backbone and year of initiation.
Results
The cohort included 390 adults; 88.7% were male, 66.4% were of white ethnicity, their median age was 40 years, there was a median of six weight measurements, 2.2 years from diagnosis to ART initiation, 2.9 years from ART to the last weight measurement, and weight and body mass index at initiation were 75 kg and 24.1 kg/m2 respectively. Of these, 254 (65%) started an INSTI. The average pre-ART rate of weight gain was 0.44 kg/year [95% confidence interval (CI): 0.19-0.70], increasing to 0.88 kg/year (0.63-1.10, p = 0.04) after ART initiation. Our adjusted model found no evidence of an association between ART regimen and rate of weight gain.
Conclusions
Weight increased in the cohort both pre- and post-ART. We found no evidence of a higher rate of weight gain following ART initiation with an INSTI compared with other regimens.
RESULTS
Of 682 identified ART-naïve individuals, 492 were started on regimens meeting the inclusion criteria; of these, 486 received this regimen for ≥ 6 months. In 424/486 participants, an undetectable VL was observed ≤ 6 months after ART initiation. A further 25 were excluded because pre- and post-ART weight measurements were outside the specified time-frame; nine were excluded due to no baseline HIV-1 VL and CD4 T-cell data. The final sample for analysis included 390 individuals with 926 pre-ART and 1415 post-ART weights. The median [interquartile range (IQR)] number of weight measurements was 2 (1-3) pre-ART and 3 (2-5) post-ART. There was a median (IQR) of 2.2 (0.3-4.7) years from HIV-1 diagnosis to ART initiation, and 2.9 (1.3-4.1) years between ART initiation and last weight.
The cohort was predominantly male (n = 346, 89%), of white ethnicity (n = 259, 66%), Centres for Disease Control category A (n = 345, 89%), with a median (IQR) age at ART initiation of 40 years (34-46), and a median pre-ART CD4 count and VL of 390 (290-528) cells/μL and 4.5 (3.9-5.1) log10copies/mL, respectively. The median (IQR) weight and BMI at ART initiation were 75.0 (68.0-83.0) kg and 24.1 (22.2-26.8) kg/m2, respectively. Of the 254 (65%) starting an INSTI, 196 (77%) started RAL and 58 (23%) DTG; 122 (48%) started TDF/FTC and 132 (52%) started ABC/3TC. None had received PrEP. Additional cohort characteristics are presented in Table S1.
The unadjusted mixed-effects model showed higher weights at initiation in the INSTI groups [NNRTI, 75.5 kg (95% CI: 73.2-77.7); PI, 73.9 kg (70.1- 77.7); DTG, 77.1 kg (74.0-80.1); RAL, 78.9 kg (76.7-81.0)] and an average pre-ART rate of weight gain of 0.44 kg/year (0.19-0.70) with an increase to 0.88 kg/year (0.63-1.10, p = 0.04) after ART initiation (Table S2). The adjusted model spline plots show trends of greater weight loss per year prior to ART initiation with lower CD4, higher baseline VL and higher age (Figure 1b,f,j; predicted average rates of change are shown for CD4, while differences in the predicted average rate of change relative to a reference value are shown for age and VL). Characteristics associated with weight loss prior to ART initiation were associated with a greater relative change in the rate of weight loss/gain at ART initiation (Figure 1d,h,l), with low CD4 and high VL at baseline associated with the larger absolute rates of weight gain on ART (Figure 1c,g).
Overall, when adjusting for the characteristics in Table 1, individuals had a higher rate of weight gain pre-ART for all other regimens compared with those starting an NNRTI, although not to a significant degree. The change in rate of weight gain after starting ART is positive where the rate of gain increased and negative where it decreased (e.g. black Africans, rate increased by 0.49 kg/year from a loss of -0.3 kg/year to a gain of 0.46 kg/year). This did not significantly differ by ART regimen relative to NNRTI [RAL, ‒0.76 kg/year (-2.36-0.84, p = 0.36); DTG, -1.01 kg/year (-2.97-0.95, p = 0.32); PI, -0.45 kg/year (-1.97-1.07, p = 0.57); NNRTI (reference); Table 1]. Similar findings were seen for BMI (Table 1, Figure S1). There was some evidence that individuals starting ART more recently had higher baseline weights [2007-2010, -0.82kg (-6.82-5.18, p = 0.79); 2011-2013, -0.13 kg (-5.29-5.03, p = 0.96); 2014-2016 (reference); 2017-2020, +2.57 kg (-0.99-6.13, p = 0.16)].
Further adjustment for patient comorbidities showed an increased rate of weight gain post-ART in those with depression (+1.29 kg/year, 95% CI: 0.32-2.26, p = 0.01) and greater weight at ART initiation for those with hypertension (+8.07 kg, 95% CI: 3.24-12.91, p = 0.001) and NAFLD (+9.48 kg, 95% CI: 5.04-13.91, p < 0.0001), without any other statistically significant findings. This did not affect the overall conclusions from the analyses regarding ART regimens.
DISCUSSION
Our real-world cohort was on average gaining weight both pre- and post-ART in our unadjusted analysis. However, we did not observe a difference in the rate of weight or BMI gain after starting an ART regimen containing RAL or DTG compared with a PI- or NNRTI-based regimen. Our study is novel in that we incorporated weights before baseline to assess pre-ART trends. We analysed the rate at which weight is gained (or lost), not the absolute weight change. Our aim was to identify whether the inclusion of INSTIs within the ART regimen was associated with an acceleration in the rate of weight gain as opposed to quantifying absolute weight change.
Observations of absolute weight gain after starting INSTIs in other treatment-naïve cohorts and a pooled analysis of randomized studies have been reported [6, 7], although contrasting findings have been reported in switch cohorts [8]. Although ART-naïve individuals may experience gains in absolute weight after starting ART, the cause is likely to be multifactorial. One hypothesis to explain the variability of effect could be that INSTIs sensitize PLWH to gaining weight, the melanocortin-4 receptor pathway being one proposed mechanism [9]. However, absolute gains are probably influenced by lifestyle [10], cultural attitudes to weight [11], the individual's perception of their weight [12], and host genetics/microbiome [13]. This may explain why weight gain is not a universal phenomenon for all individuals on ART and for all regimens [14]. Improved tolerability of modern regimens could also be a factor and is more easily identified when full adverse event data from studies are reported [7]. For individuals who undergo extremely rapid weight gain with ART, host/genetic factors, yet to be elucidated, could amplify or synergize such an effect.
Individuals with a more recent year of ART initiation were heavier at initiation. This probably reflects the fact that PLWH are now diagnosed and started on treatment earlier, with fewer patients presenting with advanced disease and the associated weight loss. However, it may also reflect the increasingly obesogenic environment that equally affects the general population; 64% of adults in England are overweight or obese (36% and 28%, respectively) [15], with a similar prevalence observed in comparable countries [16]. An increased rate of weight gain was observed in those starting treatment with lower CD4, higher VL and older age, probably associated with a return to health. Both findings have been observed in other HIV cohorts [6].
Our study is limited by being a relatively small, single-centre cohort comprised predominantly of white males. This restricts the generalisability of our findings to other groups taking INSTI-containing regimens, particularly black females who are thought to be susceptible to weight gain. Similarly, TAF was excluded due to small numbers (n = 4), precluding assessment of the potential effect of TAF as noted in other studies [7, 17]. We were also unable to include other INSTIs (cabotegravir, elvitegravir and bictegravir) due to low use.
Our study will not have captured extremely rapid changes in weight that may have occurred during the 6 months after ART initiation if this caused a regimen switch. However, only a small number of individuals were excluded on this basis (n = 6) and our primary focus was the rate of weight change across the whole cohort following a longer period of INSTI exposure. This approach allowed for sufficient time to determine if any immediate acceleration of weight gain is maintained.
CONCLUSIONS
Weight change after ART initiation is probably a complex, multifactorial process. These data suggest that RAL- and DTG-containing regimens in treatment-naïve individuals are not associated with an acceleration in weight gain compared with other ART regimens, beyond that expected following viral suppression, within the limitations specified. Research incorporating assessment of general weight gain factors such as diet and exercise would be beneficial, mitigating the under-reporting and recording bias with diet recall, although this is often logistically challenging [18]. Comorbidities also need to be considered; depression was associated with increased rate of weight gain, a condition with high prevalence [19] where both the disease and the treatment can affect weight [20].
Caution should be shown in attributing weight gain as a causal effect of INSTI use in the absence of adequate follow-up and monitoring. More data are needed to evaluate individuals switching away from INSTIs because of weight gain. This should include capturing obesogenic factors; whether the weight gain continues, plateaus or reverses; weight distribution, and whether any clinical or metabolic sequelae manifest.
|
|
| |
| |
|
|
|
