A US study of 6451 people with HIV who switched from antiretroviral drug regimens containing the older form of tenofovir, tenofovir disoproxil fumarate (TDF), to regimens containing the newer form, tenofovir alafenamide (TAF), shows rises in participants' cholesterol and triglycerides – blood lipids (fats) that raise the risk of cardiovascular disease.
The risk was moderated by the fact that levels of high-density lipoprotein (HDL, also known as ‘good’ cholesterol), which protects against heart disease, also increased. Nonetheless, there was a slight but significant rise in the ten-year risk of cardiovascular disease in patients switching from TDF to TAF.
The proportion of patients who were classed as having dyslipidaemia or severe dyslipidaemia (potentially harmful blood lipids) levels rose, in the case of total cholesterol, from 5.1% to 9.5% and the proportion who had a higher than 7.5% chance of developing cardiovascular disease (CVD) over the next ten years rose from 29% to 31%.
While this 2% increase may not be clinically significant, and the proportion at elevated risk of CVD who were being prescribed statins to lower cholesterol increased by 10% after they switched to TAF, it was still the case that 59% of those with elevated CVD risk were not receiving statins (statins do not affect triglycerides).
These findings are similar to some previously reported studies, such as an Irish study in 2020, and a US study from 2019 which found an increase of 13% in the average cardiovascular risk score after switching from TDF to TAF. But these studies only involved 164 and 110 patients, respectively. The size of the cohort in this study adds weight to the findings.
TAF was developed to avoid the high blood levels of tenofovir that lead to reduced kidney function and bone mineral density in many people who take it.
Studies have found a raised risk of kidney disease in some people on TDF, but minimal risk in younger people, including in another study of OPERA, which is the cohort analysed for the new study. They have also found low rates of side effects in people taking TDF for PrEP, probably because they are on average younger than people on treatment. But TAF offers benefits for people who already have reduced kidney function, especially older people (kidney function naturally declines with age) and one study found switching to TAF improved bone mineral density in the over-60s.
One study only found a lower rate of kidney and bone-related side effects in people taking TAF if they were also taking cobicistat or ritonavir as accessory drugs used to boost the levels of protease inhibitors, or the integrase inhibitor elvitegravir. Others have found that switching does improve markers of kidney function and bone mineral density, but that the clinical difference between the two drugs to be relatively slight.
As time has gone by, however, studies have also found that switching from TDF to TAF has some negative effects. As well as the previously mentioned studies documenting increases in lipids, some studies have found significant weight gain in patients switching to TAF. The first study that documented this was ADVANCE, which compared weight gain in three different regimens for people starting treatment. Patients who took both TAF and dolutegravir had about twice the weight gain as people starting TDF and dolutegravir. In women these gains were of clinical significance, with 51% of women who switched to dolutegravir and TAF gaining 10% or more of weight.
A 2020 study from Ireland found that TAF independently led to weight gain, and then four separate studies in 2021 confirmed this finding, with one attributing a 30% average increase in body-mass index (BMI) to TAF. A 2020 analysis of people switching from TDF to TAF in the OPERA cohort identified rapid weight gain, regardless of whether people were taking an integrase inhibitor or not.
One question yet to be answered is whether weight gains and lipid increases are due to an active property of TAF, or whether TDF reduces lipids and weight gain, with this benefit being lost when it is stopped. One meta-analysis has suggested it might, but this is difficult to disentangle from weight loss due to TDF side effects (it can cause nausea and gastrointestinal upset, primarily in the first month) or from the weight gain often seen in people who are returning to better health.
Whatever the cause, one recent study concluded that a lot of switching from TDF to TAF among PrEP users was unnecessary and some might be harmful.
The present study
Dr Laurence Brunet and colleagues from the private scientific research company Epividian selected patients from the Observational Pharmaco-epidemiology Research and Analysis (OPERA) cohort, which in total includes 93,170 people with HIV attending 84 clinics in 18 US states.
For this study, patients switching from TDF to TAF between November 2015 and March 2018 were selected. They had to have had their lipids measured no more than six months before switching, and at any time afterwards (the median time after switching was 3.9 months). People taking TDF or TAF for PrEP were not included.
Of the 6451 people included, 16% were women, 33% were Black and 37% were Hispanic. At baseline, their average age was 48 and their average CD4 count was 645; 18% were not virally suppressed. They had been on TDF for a median time of two years and five months.
All lipid levels rose after switching. Total cholesterol rose from 4.5 to 4.76 millimols per litre (mmols/l); LDL (low-density lipoprotein, 'bad' cholesterol) from 2.51 to 2.69; HDL from 1.16 to 1.19; and triglycerides from 1.43 to 1.57. The relative increases were 5.8%, 7.2%, 2.7% and 9.8% for the four classes of lipids.
These increases are expressed in European units. In the original paper the US units of milligrams per deciliter (mg/dl) were used, which also translate into different percentage increases of 7.9%, 11.1%, 7.1% and 24%, respectively. However, this makes no difference to the proportion of participants who were classed as having normal, borderline abnormal, abnormal or severely abnormal lipid levels.
Averaged over the cohort, total cholesterol and triglyceride levels were classed as normal and stayed so after switching; HDL levels were classed as borderline abnormal and stayed so; but average LDL levels, which were normal before switching, were classed as borderline abnormal afterwards.
The proportion of participants classed as having dyslipidaemia or severe dyslipidaemia is probably more important for gauging CVD risk.
- For total cholesterol, the proportion of people in these categories rose from 5% to 9.5%;
- For LDL, from 15.5% to 22%;
- For triglycerides from 21% to 27%.
- More encouragingly, the proportion who had abnormally low HDL decreased, from 35% to 30%, and this had some moderating effect on heart attack risk.
In relative terms, although only a few per cent of the cohort were classed as having severe dyslipidaemia, the proportion classed as having it more than doubled for total cholesterol from 0.7% to 1.9%, and nearly doubled for LDL, from 3.4% to 6.5%.
The researchers did a separate analysis for the 4328 people who only switched from TDF to TAF and did not change their other drugs. This found a similar but slightly larger increase in blood lipids and in the proportions of people classed as dyslipidaemic and severely dyslipidaemic.
No apparent effect of booster drugs
Because the booster drugs cobicistat or ritonavir might have an impact on side effects of TDF and TAF, the researchers also classified participants by whether they were taking boosters or not. Fifty-one per cent were taking booster drugs before the switch and 62% after.
" The study can’t tell whether the rises in lipids are due to TAF raising them or TDF lowering them."
Booster use was largely due to the fact that elvitegravir/cobicistat/emtricitabine/TDF (Stribild) or elvitegravir/cobicistat/emtricitabine/TAF (Genvoya) were the most popular single combination pills. Before the TDF/TAF switch, 27% were taking a regimen containing elvitegravir and cobicistat and after the switch, 42.5%. In contrast the use of boosted protease inhibitors (mainly darunavir) fell from 24% to 19%.
The proportions of patients who had dyslipidaemia or severe dyslipidaemia increased in those who’d had boosted regimens before and after the switch, in patients who were not boosted either before or after, and in those who were not taking boosters before, but were after.
In contrast, in the quite small number of patients (165) who were taking boosted regimens before the switch but not after (largely patients switching to dolutegravir), there was no change in the proportion who had dyslipidaemia or severe dyslipidaemia. This is the only evidence for any kind of separate influence of the booster drugs.
CVD risk, statins and conclusions
The proportion who fell into the category of having a greater than 7.5% risk of cardiovascular disease (CVD) in the next ten years increased only slightly, from 29% to 31%. Of these, 31% were taking statins to control cholesterol levels before they switched and 41% after, meaning that 59% whose levels could be controlled by statins were missing out.
In addition however, 12% of patients with low CVD risk were taking statins before the switch and 16% after, suggesting some inappropriate prescribing of statins, as well as lack of appropriate prescribing.
The researchers acknowledge that the clinical risk of switching from TDF to TAF may be small and can only be evaluated in a longer study that looks at CVD incidence. They also point out that their study can’t tell whether the rises in lipids are due to TAF raising them or TDF lowering them. This could only be settled by a larger study with arms of patients who went from regimens not containing tenofovir to ones containing TAF, and from ones containing TDF to ones not containing tenofovir.
However, this study does strengthen the evidence that, at least in a minority of patients, switching from TDF to TAF may not be risk-free, especially if their kidney function and bone mineral status do not justify it.
Brunet L et al. Switch from tenofovir disoproxil fumarate to tenofovir alafenamide in people living with HIV: lipid changes and statin underutilization. Clinical Drug Investigation, September 2021 (open access).