Bone Mineral Density, Kidney Function, Weight Gain and Insulin Resistance in Women Who Switch From TDF/FTC/NNRTI to ABC/3TC/DTG

F Ibrahim; A Samarawickrama; L Hamzah; R Vincent; Y Gilleece; L Waters; S Kegg; B Barbini; L Campbell; FA Post


HIV Medicine. 2021;22(2):83-91. 

In This Article


Women aged 40 years and over who switched from TDF/FTC/NNRTI to ABC/3TC/DTG experienced increases in BMD and CD4 cell counts, and reductions in proteinuria and plasma insulin concentrations. They also experienced weight gain, although we found no evidence that this was accompanied by reduced insulin sensitivity or metabolic syndrome. Treatment-limiting adverse events, mainly neuropsychiatric side effects including suicidality and hypersensitivity reactions, were relatively common.

As hypothesized, our switch strategy was associated with modest improvements of BMD of the hip and lumbar spine at 48 weeks. In the ACTG A5224s, predominantly male participants who initiated TDF/FTC/NNRTI (vs. ABC/3TC/EFV) experienced a mean reduction of 3.8% (vs. 2.5%) at the hip and 3.5% (vs. 1.6%) at the lumbar spine, indicating that the TDF/FTC backbone resulted in about 1.3% and 1.9% additional reduction in BMD at the hip and lumbar spine, respectively.[4] In the Gilead study 1089, participants on unboosted regimens who switched from TDF/FTC to tenofovir alafenamide (TAF)/FTC experienced increases in BMD (1.1% at the hip and 1.5% at the lumbar spine),[29] while those in Gilead study 1717 who switched from ABC/3TC to TAF/FTC had no change in BMD at either the hip or lumbar spine.[30] The change in total hip and lumbar spine BMD in our study is of the same magnitude as observed in study 1089 and a study in which participants switched from TDF/FTC/EFV to darunavir/ritonavir.[8] While many clinicians discontinue TDF in individuals found to have osteoporosis or severe osteopenia, it remains uncertain if the resulting modest improvements in BMD translate into reduced fracture risk.

The switch strategy resulted in worsening of eGFR (estimated by creatinine) and no change in eGFR (estimated by cystatin C). We observed a reduction in (total) proteinuria in those who switched to ABC/3TC/DTG, but this was not accompanied by improvement in other markers of tubular function. Total proteinuria in the 'normal' range in people with HIV consists predominantly of low-molecular-weight proteins and may thus serve as a non-specific marker of renal tubular function.[31] In the switch arm, we observed a greater reduction (43% vs. 16%) in RBPCR, a specific marker of tubule (dys)function, but this failed to reach statistical significance, perhaps because our study population, with a high proportion of black participants exposed to unboosted ART regimens, was at low risk of renal tubular dysfunction.[19,31] Fractional excretion of phosphate was not affected by switching off TDF regimens, consistent with data from an earlier study.[8]

Unlike the participants who remained on TDF/FTC/NNRTI, those in the switch arm gained an average of 1.8 kg by 48 weeks. Weight gain with integrase strand transfer inhibitors (INSTIs) in people with HIV is increasingly recognized, with a pooled analysis of eight randomized controlled clinical trials conducted in ART-naïve individuals suggesting that DTG (and bictegravir, vs. EFV)-containing regimens are associated with greater weight gain (c. 2.5 kg by 96 weeks).[32] The greatest weight gain was observed in black women; while TAF/FTC was associated with > 10% weight gain, no difference was observed between TDF/FTC and ABC/3TC. Weight gain (> 10%, vs. < 10%) was not associated with a significant increase in fasting glucose or triglycerides, or with an increased incidence of diabetes or hyperglycaemia-related adverse events (although a trend towards a higher incidence was observed in those with > 10% weight gain).[32] A large clinical trial that compared TAF/FTC/DTG with TDF/FTC/DTG and TDF/FTC/EFV in southern Africans also showed significantly greater weight gain (both lean and fat mass) in the DTG arms, especially in female participants and when DTG was co-administered with TAF/FTC. There was no difference in change in triglycerides at 48 weeks between the three study arms, and slightly smaller increases in fasting glucose in the DTG arms.[33] Our exploratory analysis, which found a reduction in plasma insulin concentrations in the ABC/3TC/DTG arm and no increase in insulin resistance or metabolic syndrome, are in agreement with these data, although larger and more prolonged studies are required to define the long-term metabolic consequences of INSTI-associated weight gain.

A slightly higher proportion (15.3%) of women than anticipated discontinued ABC/3TC/DTG due to adverse events despite screening for HLA-B5701 polymorphisms.[34] Several participants developed treatment-limiting neuropsychiatric adverse events including two with suicidality. While generally mild and no more common in randomized trials of ART-naïve individuals,[33] neuropsychiatric adverse events were more frequent in the DTG arm of one recent switch trial (5% vs. < 1%).[35] Treatment-limiting neuropsychiatric adverse events occurred in about 3.5% of DTG recipients in cohort studies,[36] more frequently with DTG (vs. other INSTI),[37–40] and with greater frequency in women, those aged 60 years and over, and DTG/ABC co-administration in one study[37] but not others.[38,39]

Women are severely underrepresented in phase III clinical trials of ART-naïve people with HIV. This study demonstrates that it is feasible to conduct interventional studies exclusively in women with unrestricted access to ART. However, this study has several limitations. We investigated a treatment switch of both the NRTI backbone (TDF/FTC to ABC/3TC) and the third agent (NNRTI to DTG) and thus are unable to dissect the effects of each of these strategies; in addition the NNRTI at baseline (e.g. EFV vs. other) may have had an impact on subsequent changes in metabolic parameters in the switch arm. A relatively large proportion of women in the ABC/3TC/DTG arm discontinued the study, resulting in loss of power; this was mitigated by utilizing multiple imputations for missing data. A proportion of women attended the baseline or week 48 visit in a non-fasting state which precluded them from being included in the analyses of insulin resistance and metabolic syndrome, and there was a substantial imbalance in weight and BMI at baseline between those who were randomized to switch to ABC/3TC/DTG vs. remain on TDF/FTC/NNRTI. The peri-menopausal status of some of the participants may have had an impact on our ability to observe ART-attributable changes in bone turnover and BMD. Finally, as we enrolled no men, women < 40 years old, and few women of non-black ethnicity (in whom peak bone mass, BMD and bone turnover in relation to the vitamin D/PTH axis may differ from black women), the results of our study may not be applicable to these populations.

In conclusion, women who switched from TDF/FTC/NNRTI to ABC/3TC/DTG experienced increases in hip and lumbar spine BMD, and weight gain that was not accompanied by worsening insulin resistance. These results need to be confirmed in larger and more diverse populations with more prolonged follow-up.