October/November 2015 

A study about switching from TDF to TAF

Researchers enrolled HIV-positive participants who were taking regimens containing TDF (tenofovir disoproxil fumarate), the original formulation of tenofovir. All participants had been on their current regimen for at least 96 weeks and had a viral load less than 50 copies/ml during this time. Once in the study, participants were randomly assigned to either receive a regimen containing TAF (tenofovir alafenamide), the new formulation of tenofovir, or to continue with their existing TDF-containing regimen. Researchers presented interim results analysed from 48 weeks of data (the study is planned to continue for 96 weeks). The results suggest that switching to a TAF-based regimen is generally safer than continuing to take TDF-containing regimens, particularly for bone and kidney health.

Study details

Prior to being randomized, participants were on the following regimens, all of which contained TDF:

  • Stribild, a fixed-dose combination of  TDF + FTC + elvitegravir + cobicistat (459 people)
  • TDF + FTC + atazanavir + ritonavir  (601 people)
  • Atripla, a fixed-dose combination of TDF + FTC + efavirenz (376 people)

Participants were randomized on a 2:1 ratio to receive one of the following regimens:

  • TAF + FTC + elvitegravir + cobicistat (959 people)
  • continue their existing regimen (477 people)

The average profile of participants upon entering the study was as follows:

  • age – 41 years
  • 89% men, 11% women
  • CD4+ cell count – 670 cells/mm3
  • proportion who had less than 200 CD4+ cells – less than 1%
  • proportion who had mild to moderate levels of sugar in their urine – 9% (suggestive of kidney dysfunction)


Proportion of participants with a viral load less than 50 copies/ml at week 48:

  • TAF-based regimen – 97%
  • TDF-based regimen – 93%

The proportion of participants with virological failure was 1% in each of the randomized interventions.

The proportion of participants with no virological data at week 48 was as follows:

  • TAF-based regimen – 2%
  • TDF-based regimen – 6%

A note about claims of superiority

The overall statistical analysis suggests that a TAF-based regimen is statistically superior to a TDF-based regimen. However, this statistical difference arose because more participants taking a TDF-based regimen did not have data available for analysis at week 48. This may have occurred because they dropped out of the study or were not able to be found by researchers or there may be other reasons that the data were missing. Furthermore, the regimen that TAF users received contained the integrase inhibitor elvitegravir. Integrase inhibitors are the most potent anti-HIV drugs; they quickly reduce viral load and an integrase inhibitor-based regimen would likely outperform a regimen containing a protease inhibitor (such as atazanavir) or a non-nuke (such as efavirenz). Therefore, the claim of statistical superiority may be technically correct, but it might also arise because of factors likely unrelated to the use of TAF.

Comparing regimens

At week 48, the following proportion of participants had a viral load that was less than 50 copies/ml:

  • prior use of Atripla and then switched to a TAF-based regimen – 96%
  • prior use of Atripla and continued on that regimen – 90%
  • prior use of atazanavir and then switched to a TAF-based regimen – 97%
  • prior use of atazanavir and continued on that regimen – 92%
  • prior use of Stribild and then switched to a TAF-based regimen – 98%
  • prior use of Stribild and continued on that regimen – 97%

Side effects and complications

Overall, the proportion of participants who left the study prematurely due to side effects was as follows:

  • TAF-based regimen – 1%
  • TDF-based regimen – 3%

Some of these people left because of kidney-related events, as follows:

  • TAF-based regimen – one case each of kidney failure and declining kidney function
  • TDF-based regimen – one case of chronic kidney disease and a handful of cases of kidney injury

Other reasons that participants receiving either TAF or TDF gave for leaving prematurely seemed to be related to anxiety and depression.

General side effects

Here is the distribution of side effects reported by at least 5% of participants, showing marginal differences between the two formulations of tenofovir:


  • TAF users – 10%
  • TDF users – 9%


  • TAF users – 7%
  • TDF users – 4%

Bone/joint pain

  • TAF users – 6%
  • TDF users – 5%

Problems falling asleep or staying asleep

  • TAF users – 5%
  • TDF users – 6%

Back pain

  • TAF users – 5%
  • TDF users – 5%


  • TAF users – 5%
  • TDF users – 3%

Abnormal lab test results

Moderate to seriously abnormal lab test results occurred in 25% to 30% of participants, depending on the medicines they were taking. As with other studies of TAF, this drug did not appear to cause more abnormal lab test results than TDF.

The most common abnormal blood test results that were moderate to serious in severity were distributed as follows:

Elevations in the enzyme creatine kinase (possibly suggestive of muscle injury)

  • TAF users – 10%
  • TDF users – 10%

Elevations in the liver enzyme AST (suggestive of liver injury)

  • TAF users – 5%
  • TDF users – 7%

Elevations in the liver enzyme ALT (suggestive of liver injury)

  • TAF users – 5%
  • TDF users – 5%

Less-than-normal levels of neutrophils (these cells are part of the immune system)

  • TAF users – 4%
  • TDF users – 3%

Less-than-normal levels of phosphate (suggestive of kidney injury)

  • TAF users – 2%
  • TDF users – 3%

Focus on lipids (cholesterol and triglycerides)

In general, when researchers assessed blood samples that were taken when participants had been fasting, they found that, over time, there were modest elevations in lipid levels among TAF users compared to TDF users. The lipids assessed included the following:

  • total cholesterol
  • bad cholesterol (LDL-C)
  • good cholesterol (HDL-C)
  • triglycerides

However, the ratio of total cholesterol to HDL-C was similar in TAF and TDF users, suggesting that their future risk for cardiovascular disease was more or less the same in people who received TAF or TDF.

The distribution of participants who initiated lipid-lowering therapy during the study was as follows:

  • TAF users – 8%
  • TDF users – 6%

Comparing changes in bone density

Overall, a statistically significant increase in bone density (2%) occurred among TAF users. In contrast, participants who remained on TDF had their bone density decrease slightly.

Changes in osteoporosis status

Bone thinness can be grouped into two categories—osteopenia, a mild form of decreased bone density, and osteoporosis, a severe form of bone density loss.

Looking at the spine

At the start of the study, osteopenia and osteoporosis were distributed among participants as follows:

TAF-based regimen

  • 36% had osteopenia in the spine
  • 6% had osteoporosis of the spine

TDF-based regimen

  • 35% had osteopenia in the spine
  • 7% had osteoporosis of the spine

After 48 weeks, the distribution of these conditions in the spine was as follows:

TAF-based regimen

  • 32% had osteopenia
  • 5% had osteoporosis

TDF-based regimen

  • 37% had osteopenia
  • 8% had osteoporosis

Looking at the hips

When researchers assessed bone density in the hips, the distribution of osteopenia and osteoporosis at the start of the study was as follows:

TAF-based regimen

  • 31% had osteopenia in the hips
  • 0.7% had osteoporosis of the hips

TDF-based regimen

  • 32% had osteopenia in the hips
  • 1.3% had osteoporosis of the hips

After 48 weeks, the distribution of osteoporosis and osteopenia in the hips was as follows:

TAF-based regimen

  • 26% had osteopenia
  • 0.7% had osteoporosis

TDF-based regimen

  • 32% had osteopenia
  • 2.1% had osteoporosis

Kidney injury

Research has found that the original formulation of tenofovir (TDF) can, in some users, cause varying degrees of kidney injury. Therefore, in clinical trials of TDF (and its successor compound, TAF) it is important to conduct detailed and complex assessments of kidney health.

In the present study, in general, participants taking TDF-based regimens were more likely to develop signals of kidney injury than participants on TAF-based regimens.

Researchers assessed certain proteins in the urine as follows:

  • protein to creatinine ratio
  • albumin to creatinine ratio
  • retinol-binding protein to creatinine ratio
  • beta2-microglobulin to creatinine ratio

Note: Some of the tests above, particularly concerning retinol-binding protein and beta2-microglobulin, are usually only done as part of research, not routine care in the clinic.

While participants were on a TAF-based regimen, levels of the four proteins in the urine fell, suggesting improved kidney health in participants. Another measure of kidney health, eGFR (estimated glomerular filtration rate), improved very modestly in TAF users. In contrast, changes to the urinary proteins among participants on a TDF-based regimen were unfavourable.

Key points

Switching from a TDF-based regimen to a TAF-based regimen appears to be safer for kidney and bones.

TAF appears to be equivalent to TDF when used as part of combination anti-HIV therapy.

—Sean R. Hosein


Mils A, Andrade-Villanueva J, DiPerri G, et al. Switching from a tenofovir disoproxil fumarate (TDF)-based regimen to a tenofovir alafenamide (TAF)-based regimen: data in virologically suppressed adults through week 48 of treatment. In: Program and abstracts of the 8th IAS Conference on HIV Pathogenesis, Treatment and Prevention, 19-22 July 2015, Vancouver, Canada. Abstract TUAB0102.